add claude-obsidian
This commit is contained in:
김경종
@@ -0,0 +1,37 @@
|
||||
{
|
||||
"nodes":[
|
||||
{"id":"title","type":"text","text":"# 🧠 Cosmic Brain\n### Knowledge Map — Fibonacci Spiral\n*12 nodes at golden angle 137.5° — the sunflower pattern*","x":-280,"y":-301,"width":560,"height":141},
|
||||
{"id":"c-idx","type":"file","file":"wiki/concepts/_index.md","x":120,"y":-40,"width":270,"height":80,"color":"2"},
|
||||
{"id":"e-idx","type":"file","file":"wiki/entities/_index.md","x":-411,"y":209,"width":270,"height":80,"color":"6"},
|
||||
{"id":"s-idx","type":"file","file":"wiki/sources/_index.md","x":-101,"y":-488,"width":270,"height":80,"color":"4"},
|
||||
{"id":"dash","type":"file","file":"wiki/meta/dashboard.md","x":187,"y":372,"width":260,"height":108,"color":"5"},
|
||||
{"id":"ov","type":"file","file":"wiki/overview.md","x":-702,"y":-141,"width":260,"height":80,"color":"5"},
|
||||
{"id":"llm-wiki","type":"file","file":"wiki/concepts/LLM Wiki Pattern.md","x":401,"y":-381,"width":260,"height":80,"color":"2"},
|
||||
{"id":"hot-c","type":"file","file":"wiki/concepts/Hot Cache.md","x":-313,"y":623,"width":260,"height":80,"color":"2"},
|
||||
{"id":"hot-m","type":"file","file":"wiki/hot.md","x":-469,"y":-691,"width":260,"height":80,"color":"5"},
|
||||
{"id":"comp","type":"file","file":"wiki/concepts/Compounding Knowledge.md","x":593,"y":227,"width":280,"height":80,"color":"2"},
|
||||
{"id":"karp","type":"file","file":"wiki/entities/Andrej Karpathy.md","x":-891,"y":275,"width":260,"height":80,"color":"6"},
|
||||
{"id":"log-m","type":"file","file":"wiki/log.md","x":235,"y":-822,"width":260,"height":80,"color":"5"},
|
||||
{"id":"lbl-concepts","type":"text","text":"**💡 CONCEPTS**","x":420,"y":-460,"width":160,"height":40,"color":"2"},
|
||||
{"id":"lbl-entities","type":"text","text":"**👤 ENTITIES**","x":-960,"y":195,"width":160,"height":40,"color":"6"},
|
||||
{"id":"lbl-sources","type":"text","text":"**📄 SOURCES**","x":-200,"y":-580,"width":160,"height":40,"color":"4"},
|
||||
{"id":"lbl-nav","type":"text","text":"**🗂️ NAVIGATION**","x":-560,"y":-790,"width":190,"height":40,"color":"5"},
|
||||
{"id":"hub","type":"file","file":"wiki/index.md","x":-230,"y":-141,"width":350,"height":2095,"color":"5"}
|
||||
],
|
||||
"edges":[
|
||||
{"id":"e-hub-cidx","fromNode":"hub","fromSide":"right","toNode":"c-idx","toSide":"left","label":"concepts"},
|
||||
{"id":"e-hub-eidx","fromNode":"hub","fromSide":"bottom","toNode":"e-idx","toSide":"top","label":"entities"},
|
||||
{"id":"e-hub-sidx","fromNode":"hub","fromSide":"top","toNode":"s-idx","toSide":"bottom","label":"sources"},
|
||||
{"id":"e-hub-dash","fromNode":"hub","fromSide":"right","toNode":"dash","toSide":"left"},
|
||||
{"id":"e-hub-ov","fromNode":"hub","fromSide":"left","toNode":"ov","toSide":"right"},
|
||||
{"id":"e-hub-hot","fromNode":"hub","fromSide":"left","toNode":"hot-m","toSide":"right"},
|
||||
{"id":"e-hub-log","fromNode":"hub","fromSide":"top","toNode":"log-m","toSide":"bottom"},
|
||||
{"id":"e-cidx-llm","fromNode":"c-idx","fromSide":"right","toNode":"llm-wiki","toSide":"left"},
|
||||
{"id":"e-cidx-hc","fromNode":"c-idx","fromSide":"bottom","toNode":"hot-c","toSide":"top"},
|
||||
{"id":"e-cidx-comp","fromNode":"c-idx","fromSide":"right","toNode":"comp","toSide":"left"},
|
||||
{"id":"e-eidx-karp","fromNode":"e-idx","fromSide":"left","toNode":"karp","toSide":"right"},
|
||||
{"id":"e-llm-karp","fromNode":"llm-wiki","fromSide":"bottom","toNode":"karp","toSide":"top","toEnd":"none","color":"6","label":"originated by"},
|
||||
{"id":"e-hc-hot","fromNode":"hot-c","fromSide":"left","toNode":"hot-m","toSide":"right","label":"implemented as"},
|
||||
{"id":"e-comp-karp","fromNode":"comp","fromSide":"left","toNode":"karp","toSide":"right","toEnd":"none","color":"6"}
|
||||
]
|
||||
}
|
||||
@@ -0,0 +1,293 @@
|
||||
{
|
||||
"nodes": [
|
||||
{
|
||||
"id": "title-hero-1744041601",
|
||||
"type": "text",
|
||||
"text": "# claude-obsidian\n\n**Claude Code × Obsidian** — Persistent, compounding knowledge base.\nEvery source you add makes the wiki smarter. No vector DB. No embeddings. Just markdown.\n\n*Skills: /wiki · /wiki-ingest · /wiki-query · /wiki-lint · /autoresearch · /canvas*",
|
||||
"x": -960, "y": -3420, "width": 600, "height": 200, "color": "6"
|
||||
},
|
||||
{
|
||||
"id": "img-cover-16x9-1744041602",
|
||||
"type": "file",
|
||||
"file": "wiki/meta/claude-obsidian-cover-16x9.png",
|
||||
"x": -330, "y": -3420, "width": 420, "height": 236
|
||||
},
|
||||
{
|
||||
"id": "img-gif-cover-1744041603",
|
||||
"type": "file",
|
||||
"file": "wiki/meta/claude-obsidian-gif-cover-16x9.gif",
|
||||
"x": 130, "y": -3420, "width": 420, "height": 236
|
||||
},
|
||||
{
|
||||
"id": "img-gif-1x1-1744041604",
|
||||
"type": "file",
|
||||
"file": "wiki/meta/claude-obsidian-gif-1x1.gif",
|
||||
"x": 590, "y": -3420, "width": 200, "height": 200
|
||||
},
|
||||
{
|
||||
"id": "zone-hero-1744041605",
|
||||
"type": "group",
|
||||
"label": "Hero",
|
||||
"x": -980, "y": -3440, "width": 1820, "height": 280, "color": "6"
|
||||
},
|
||||
|
||||
{
|
||||
"id": "zone-overview-1744041606",
|
||||
"type": "group",
|
||||
"label": "What Is claude-obsidian",
|
||||
"x": -980, "y": -3110, "width": 1820, "height": 420, "color": "5"
|
||||
},
|
||||
{
|
||||
"id": "text-what-1744041607",
|
||||
"type": "text",
|
||||
"text": "## What Is It?\n\nAn LLM wiki that **compounds** with every source you add. Originated by Andrej Karpathy.\n\nInstead of re-deriving from raw documents on every query (RAG), Claude reads each source, extracts what matters, and integrates it into a structured wiki — updating entity pages, noting contradictions, strengthening the synthesis.",
|
||||
"x": -960, "y": -3090, "width": 520, "height": 200, "color": "5"
|
||||
},
|
||||
{
|
||||
"id": "text-layers-1744041608",
|
||||
"type": "text",
|
||||
"text": "## Three Layers\n\n```\n.raw/ Layer 1 — immutable sources\nwiki/ Layer 2 — LLM knowledge base\nCLAUDE.md Layer 3 — maintenance schema\n```\n\nThe LLM owns Layer 2 entirely. It creates pages, updates them, maintains cross-references, and keeps everything consistent.",
|
||||
"x": -400, "y": -3090, "width": 460, "height": 200, "color": "5"
|
||||
},
|
||||
{
|
||||
"id": "text-why-1744041609",
|
||||
"type": "text",
|
||||
"text": "## Why It Works\n\nLLMs don't get bored.\n\nThe tedious part of maintaining a wiki is bookkeeping: cross-references, contradictions, stale summaries. Humans abandon wikis because maintenance cost grows faster than value.\n\n**With Claude: maintenance cost is near zero.**\nKnowledge compounds over time.",
|
||||
"x": 100, "y": -3090, "width": 420, "height": 200, "color": "5"
|
||||
},
|
||||
{
|
||||
"id": "note-llm-wiki-pattern-1744041610",
|
||||
"type": "file",
|
||||
"file": "wiki/concepts/LLM Wiki Pattern.md",
|
||||
"x": 560, "y": -3090, "width": 240, "height": 100
|
||||
},
|
||||
{
|
||||
"id": "note-compounding-1744041611",
|
||||
"type": "file",
|
||||
"file": "wiki/concepts/Compounding Knowledge.md",
|
||||
"x": 560, "y": -2960, "width": 240, "height": 100
|
||||
},
|
||||
{
|
||||
"id": "note-hot-cache-1744041612",
|
||||
"type": "file",
|
||||
"file": "wiki/concepts/Hot Cache.md",
|
||||
"x": 560, "y": -2830, "width": 240, "height": 100
|
||||
},
|
||||
|
||||
{
|
||||
"id": "zone-skills-1744041613",
|
||||
"type": "group",
|
||||
"label": "Skills — Six Sub-Commands",
|
||||
"x": -980, "y": -2640, "width": 1820, "height": 600, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "text-skills-header-1744041614",
|
||||
"type": "text",
|
||||
"text": "## Six Skills\n\n`/wiki` · `/wiki-ingest` · `/wiki-query` · `/wiki-lint` · `/autoresearch` · `/canvas`\n\nEach skill is a focused operation on the vault. Together they form a full knowledge lifecycle.",
|
||||
"x": -960, "y": -2620, "width": 1780, "height": 80, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "img-wiki-ingest-1744041615",
|
||||
"type": "file",
|
||||
"file": "_attachments/images/canvas/skill-wiki-ingest.png",
|
||||
"x": -960, "y": -2510, "width": 400, "height": 225
|
||||
},
|
||||
{
|
||||
"id": "img-wiki-query-1744041616",
|
||||
"type": "file",
|
||||
"file": "_attachments/images/canvas/skill-wiki-query.png",
|
||||
"x": -520, "y": -2510, "width": 400, "height": 225
|
||||
},
|
||||
{
|
||||
"id": "img-autoresearch-1744041617",
|
||||
"type": "file",
|
||||
"file": "_attachments/images/canvas/skill-autoresearch.png",
|
||||
"x": -80, "y": -2510, "width": 400, "height": 225
|
||||
},
|
||||
{
|
||||
"id": "img-wiki-canvas-1744041618",
|
||||
"type": "file",
|
||||
"file": "_attachments/images/canvas/skill-wiki-canvas.png",
|
||||
"x": 360, "y": -2510, "width": 400, "height": 225
|
||||
},
|
||||
{
|
||||
"id": "text-ingest-desc-1744041619",
|
||||
"type": "text",
|
||||
"text": "**WIKI INGEST**\nDrop a source in `.raw/`, Claude reads it, extracts entities + concepts, updates 8–15 wiki pages, logs the ingest.",
|
||||
"x": -960, "y": -2255, "width": 400, "height": 100
|
||||
},
|
||||
{
|
||||
"id": "text-query-desc-1744041620",
|
||||
"type": "text",
|
||||
"text": "**WIKI QUERY**\nAsk any question. Claude reads the index, finds relevant pages, synthesizes with citations. Good answers get filed back.",
|
||||
"x": -520, "y": -2255, "width": 400, "height": 100
|
||||
},
|
||||
{
|
||||
"id": "text-autoresearch-desc-1744041621",
|
||||
"type": "text",
|
||||
"text": "**AUTORESEARCH**\nAutonomous web research loop. Searches, fetches, synthesizes, files. Based on Karpathy's autoresearch pattern.",
|
||||
"x": -80, "y": -2255, "width": 400, "height": 100
|
||||
},
|
||||
{
|
||||
"id": "text-canvas-desc-1744041622",
|
||||
"type": "text",
|
||||
"text": "**WIKI CANVAS**\nVisual layer. Add images, PDFs, and wiki notes to Obsidian canvas files with auto-positioning and zone management.",
|
||||
"x": 360, "y": -2255, "width": 400, "height": 100
|
||||
},
|
||||
|
||||
{
|
||||
"id": "zone-architecture-1744041623",
|
||||
"type": "group",
|
||||
"label": "Architecture & Workflow",
|
||||
"x": -980, "y": -2090, "width": 1820, "height": 500, "color": "2"
|
||||
},
|
||||
{
|
||||
"id": "text-arch-header-1744041624",
|
||||
"type": "text",
|
||||
"text": "## Lifecycle Flow\n\n```\nDrop source → /wiki-ingest → Wiki updated\n ↓\n /wiki-lint ←→ /wiki-query\n ↓\n /autoresearch\n ↓\n /canvas (visual)\n```",
|
||||
"x": -960, "y": -2070, "width": 560, "height": 260, "color": "2"
|
||||
},
|
||||
{
|
||||
"id": "img-workflow-loop-1744041625",
|
||||
"type": "file",
|
||||
"file": "wiki/meta/workflow-loop.gif",
|
||||
"x": -360, "y": -2070, "width": 420, "height": 236
|
||||
},
|
||||
{
|
||||
"id": "text-hot-cache-1744041626",
|
||||
"type": "text",
|
||||
"text": "## Hot Cache\n\n`wiki/hot.md` — a ~500-word context file updated after every ingest and session.\n\nLoaded automatically at conversation start. Gives Claude instant context without reading the full index.\n\n**Pattern:** hot → index → domain → page",
|
||||
"x": 100, "y": -2070, "width": 420, "height": 200, "color": "2"
|
||||
},
|
||||
{
|
||||
"id": "img-graph-view-1744041627",
|
||||
"type": "file",
|
||||
"file": "wiki/meta/image-example-graph-view.png",
|
||||
"x": 560, "y": -2070, "width": 420, "height": 236
|
||||
},
|
||||
{
|
||||
"id": "text-arch-stats-1744041628",
|
||||
"type": "text",
|
||||
"text": "**One source → 8–15 wiki pages touched**\n\nEntities, concepts, comparisons, questions — all updated in a single ingest pass.",
|
||||
"x": -960, "y": -1780, "width": 500, "height": 80
|
||||
},
|
||||
{
|
||||
"id": "img-wiki-graph-grow-1744041629",
|
||||
"type": "file",
|
||||
"file": "wiki/meta/wiki-graph-grow.gif",
|
||||
"x": -420, "y": -1780, "width": 420, "height": 236
|
||||
},
|
||||
|
||||
{
|
||||
"id": "zone-comparison-1744041630",
|
||||
"type": "group",
|
||||
"label": "Wiki vs RAG — When to Use Which",
|
||||
"x": -980, "y": -1490, "width": 1820, "height": 440, "color": "1"
|
||||
},
|
||||
{
|
||||
"id": "img-wiki-vs-rag-1744041631",
|
||||
"type": "file",
|
||||
"file": "_attachments/images/canvas/wiki-vs-rag-chart.svg",
|
||||
"x": -960, "y": -1470, "width": 800, "height": 420
|
||||
},
|
||||
{
|
||||
"id": "text-comparison-table-1744041632",
|
||||
"type": "text",
|
||||
"text": "## When Wiki Wins\n\n| Metric | LLM Wiki | RAG |\n|---|---|---|\n| Infrastructure | Markdown files | Vector DB |\n| Cost | Tokens only | Compute + storage |\n| Maintenance | Run a lint | Re-embed |\n| Scale limit | ~1000 pages | Millions |\n| Compounding | ✅ Yes | ❌ No |\n\n**Verdict:** Wiki wins at <1000 pages.\nFor millions of documents, use RAG.",
|
||||
"x": -120, "y": -1470, "width": 480, "height": 360
|
||||
},
|
||||
{
|
||||
"id": "note-wiki-vs-rag-1744041633",
|
||||
"type": "file",
|
||||
"file": "wiki/comparisons/Wiki vs RAG.md",
|
||||
"x": 400, "y": -1470, "width": 380, "height": 100
|
||||
},
|
||||
|
||||
{
|
||||
"id": "zone-demos-1744041634",
|
||||
"type": "group",
|
||||
"label": "Visual Demos",
|
||||
"x": -980, "y": -990, "width": 1820, "height": 520, "color": "3"
|
||||
},
|
||||
{
|
||||
"id": "text-demos-header-1744041635",
|
||||
"type": "text",
|
||||
"text": "## See It In Action",
|
||||
"x": -960, "y": -970, "width": 300, "height": 50, "color": "3"
|
||||
},
|
||||
{
|
||||
"id": "img-welcome-canvas-1744041636",
|
||||
"type": "file",
|
||||
"file": "wiki/meta/welcome-canvas.gif",
|
||||
"x": -960, "y": -900, "width": 420, "height": 236
|
||||
},
|
||||
{
|
||||
"id": "img-wiki-map-1744041637",
|
||||
"type": "file",
|
||||
"file": "wiki/meta/image-example-wiki-map-view.png",
|
||||
"x": -500, "y": -900, "width": 420, "height": 236
|
||||
},
|
||||
{
|
||||
"id": "img-cover-square-1744041638",
|
||||
"type": "file",
|
||||
"file": "wiki/meta/claude-obsidian-cover-square.png",
|
||||
"x": -40, "y": -900, "width": 280, "height": 280
|
||||
},
|
||||
{
|
||||
"id": "text-install-1744041639",
|
||||
"type": "text",
|
||||
"text": "## Install\n\n```bash\nclaude plugin marketplace add \\\n AI-Marketing-Hub/claude-obsidian\n\nclaude plugin install claude-obsidian\n```\n\nRepo: `github.com/AI-Marketing-Hub/claude-obsidian`",
|
||||
"x": 280, "y": -900, "width": 480, "height": 200, "color": "3"
|
||||
},
|
||||
{
|
||||
"id": "text-canvas-label-1744041640",
|
||||
"type": "text",
|
||||
"text": "Canvas Demo",
|
||||
"x": -960, "y": -630, "width": 420, "height": 40
|
||||
},
|
||||
{
|
||||
"id": "text-wikimap-label-1744041641",
|
||||
"type": "text",
|
||||
"text": "Wiki Map View",
|
||||
"x": -500, "y": -630, "width": 420, "height": 40
|
||||
},
|
||||
{
|
||||
"id": "text-cover-label-1744041642",
|
||||
"type": "text",
|
||||
"text": "Plugin Cover",
|
||||
"x": -40, "y": -600, "width": 280, "height": 40
|
||||
}
|
||||
],
|
||||
"edges": [
|
||||
{
|
||||
"id": "e-ingest-query-1744041650",
|
||||
"fromNode": "img-wiki-ingest-1744041615",
|
||||
"fromSide": "right",
|
||||
"toNode": "img-wiki-query-1744041616",
|
||||
"toSide": "left",
|
||||
"toEnd": "arrow",
|
||||
"label": "builds",
|
||||
"color": "4"
|
||||
},
|
||||
{
|
||||
"id": "e-query-auto-1744041651",
|
||||
"fromNode": "img-wiki-query-1744041616",
|
||||
"fromSide": "right",
|
||||
"toNode": "img-autoresearch-1744041617",
|
||||
"toSide": "left",
|
||||
"toEnd": "arrow",
|
||||
"label": "grows",
|
||||
"color": "4"
|
||||
},
|
||||
{
|
||||
"id": "e-auto-canvas-1744041652",
|
||||
"fromNode": "img-autoresearch-1744041617",
|
||||
"fromSide": "right",
|
||||
"toNode": "img-wiki-canvas-1744041618",
|
||||
"toSide": "left",
|
||||
"toEnd": "arrow",
|
||||
"label": "visualizes",
|
||||
"color": "4"
|
||||
}
|
||||
]
|
||||
}
|
||||
@@ -0,0 +1,72 @@
|
||||
{
|
||||
"nodes": [
|
||||
{
|
||||
"id": "zone-wiki-pages-1744033999",
|
||||
"type": "group",
|
||||
"x": -400,
|
||||
"y": 320,
|
||||
"width": 1000,
|
||||
"height": 320,
|
||||
"color": "2",
|
||||
"label": "Wiki Pages"
|
||||
},
|
||||
{
|
||||
"id": "zone-default",
|
||||
"type": "group",
|
||||
"x": -400,
|
||||
"y": -140,
|
||||
"width": 800,
|
||||
"height": 400,
|
||||
"color": "4",
|
||||
"label": "General"
|
||||
},
|
||||
{
|
||||
"id": "note-llm-wiki-1744033901",
|
||||
"type": "file",
|
||||
"file": "wiki/concepts/LLM Wiki Pattern.md",
|
||||
"x": 80,
|
||||
"y": -120,
|
||||
"width": 300,
|
||||
"height": 100
|
||||
},
|
||||
{
|
||||
"id": "text-insight-1744033945",
|
||||
"type": "text",
|
||||
"text": "## Key insight\n\nKnowledge compounds like interest.\nEvery ingest enriches 8–15 pages.",
|
||||
"x": 80,
|
||||
"y": 0,
|
||||
"width": 300,
|
||||
"height": 220,
|
||||
"color": "5"
|
||||
},
|
||||
{
|
||||
"id": "484d32f71e442aa0",
|
||||
"type": "file",
|
||||
"file": "Cosmic Brain Cover.png",
|
||||
"x": -360,
|
||||
"y": 360,
|
||||
"width": 400,
|
||||
"height": 206
|
||||
},
|
||||
{
|
||||
"id": "a8f6d861b3c08664",
|
||||
"type": "file",
|
||||
"file": "Cosmic Brain Clean.gif",
|
||||
"x": 80,
|
||||
"y": 360,
|
||||
"width": 400,
|
||||
"height": 206
|
||||
},
|
||||
{
|
||||
"id": "title",
|
||||
"type": "text",
|
||||
"text": "# Visual Reference\n\nDrop images, PDFs, and notes here.",
|
||||
"x": -440,
|
||||
"y": -360,
|
||||
"width": 400,
|
||||
"height": 120,
|
||||
"color": "6"
|
||||
}
|
||||
],
|
||||
"edges": []
|
||||
}
|
||||
@@ -0,0 +1,209 @@
|
||||
{
|
||||
"nodes": [
|
||||
{
|
||||
"id": "zone-input",
|
||||
"type": "group",
|
||||
"x": -680,
|
||||
"y": -360,
|
||||
"width": 560,
|
||||
"height": 560,
|
||||
"color": "4",
|
||||
"label": "1 · Drop Your Sources"
|
||||
},
|
||||
{
|
||||
"id": "zone-output",
|
||||
"type": "group",
|
||||
"x": 480,
|
||||
"y": -360,
|
||||
"width": 560,
|
||||
"height": 560,
|
||||
"color": "6",
|
||||
"label": "3 · Your Brain (Vault) Grows"
|
||||
},
|
||||
{
|
||||
"id": "zone-process",
|
||||
"type": "group",
|
||||
"x": -60,
|
||||
"y": -360,
|
||||
"width": 480,
|
||||
"height": 560,
|
||||
"color": "2",
|
||||
"label": "2 · I Process Everything"
|
||||
},
|
||||
{
|
||||
"id": "gif-workflow",
|
||||
"type": "file",
|
||||
"file": "wiki/meta/workflow-loop.gif",
|
||||
"x": -40,
|
||||
"y": -340,
|
||||
"width": 440,
|
||||
"height": 440
|
||||
},
|
||||
{
|
||||
"id": "main-title",
|
||||
"type": "text",
|
||||
"text": "# 🧠 Claude Obsidian\n### Drop anything. I'll build your knowledge base.",
|
||||
"x": -340,
|
||||
"y": -620,
|
||||
"width": 840,
|
||||
"height": 110
|
||||
},
|
||||
{
|
||||
"id": "sub-title",
|
||||
"type": "text",
|
||||
"text": "Images · PDFs · Markdown · Transcripts · URLs — drag, drop, or paste. I handle the rest.",
|
||||
"x": -340,
|
||||
"y": -500,
|
||||
"width": 840,
|
||||
"height": 54,
|
||||
"color": "5"
|
||||
},
|
||||
{
|
||||
"id": "text-drop-images",
|
||||
"type": "text",
|
||||
"text": "## 📸 Paste any image here\n\nHere you can paste any images and I will take care of the rest.",
|
||||
"x": -660,
|
||||
"y": -340,
|
||||
"width": 520,
|
||||
"height": 110,
|
||||
"color": "4"
|
||||
},
|
||||
{
|
||||
"id": "text-drop-pdf",
|
||||
"type": "text",
|
||||
"text": "## 📄 Or drop PDF & Markdown files\n\nJust easy drag & drop — I'll extract everything.",
|
||||
"x": -660,
|
||||
"y": 60,
|
||||
"width": 520,
|
||||
"height": 100,
|
||||
"color": "4"
|
||||
},
|
||||
{
|
||||
"id": "72b304484d8afa75",
|
||||
"type": "file",
|
||||
"file": "Cosmic Brain Clean.gif",
|
||||
"x": -668,
|
||||
"y": -216,
|
||||
"width": 536,
|
||||
"height": 276
|
||||
},
|
||||
{
|
||||
"id": "text-process",
|
||||
"type": "text",
|
||||
"text": "Reads sources → extracts entities & concepts → cross-references everything → files it all automatically.",
|
||||
"x": -40,
|
||||
"y": 110,
|
||||
"width": 440,
|
||||
"height": 80,
|
||||
"color": "2"
|
||||
},
|
||||
{
|
||||
"id": "callout-1",
|
||||
"type": "text",
|
||||
"text": "💡 **8–15 wiki pages**\nper source ingested",
|
||||
"x": -680,
|
||||
"y": 440,
|
||||
"width": 260,
|
||||
"height": 80,
|
||||
"color": "4"
|
||||
},
|
||||
{
|
||||
"id": "callout-2",
|
||||
"type": "text",
|
||||
"text": "⚡ **Hot cache**\ninstant session context",
|
||||
"x": -400,
|
||||
"y": 440,
|
||||
"width": 260,
|
||||
"height": 80,
|
||||
"color": "2"
|
||||
},
|
||||
{
|
||||
"id": "callout-3",
|
||||
"type": "text",
|
||||
"text": "🔍 **Query anything**\nyou've ever added",
|
||||
"x": -120,
|
||||
"y": 440,
|
||||
"width": 260,
|
||||
"height": 80,
|
||||
"color": "5"
|
||||
},
|
||||
{
|
||||
"id": "callout-4",
|
||||
"type": "text",
|
||||
"text": "🕸️ **Visual map**\nFibonacci graph layout",
|
||||
"x": 160,
|
||||
"y": 440,
|
||||
"width": 260,
|
||||
"height": 80,
|
||||
"color": "6"
|
||||
},
|
||||
{
|
||||
"id": "text-brain",
|
||||
"type": "text",
|
||||
"text": "## 🧠 I'll create a full map/brain for you\n\nEvery source compounds the knowledge base.",
|
||||
"x": 500,
|
||||
"y": -340,
|
||||
"width": 520,
|
||||
"height": 100,
|
||||
"color": "6"
|
||||
},
|
||||
{
|
||||
"id": "text-result",
|
||||
"type": "text",
|
||||
"text": "Cross-referenced. Searchable. Gets richer every session.",
|
||||
"x": 500,
|
||||
"y": 200,
|
||||
"width": 520,
|
||||
"height": 50,
|
||||
"color": "6"
|
||||
},
|
||||
{
|
||||
"id": "gif-graph",
|
||||
"type": "file",
|
||||
"file": "wiki/meta/wiki-graph-grow.gif",
|
||||
"x": 568,
|
||||
"y": -225,
|
||||
"width": 385,
|
||||
"height": 385
|
||||
},
|
||||
{
|
||||
"id": "cta-start",
|
||||
"type": "text",
|
||||
"text": "## 👉 Start here\n\n1. Read [[getting-started]] for a quick walkthrough\n2. Run `/wiki` in Claude Code to scaffold your vault\n3. Drop your first source into `.raw/` and say `ingest [filename]`",
|
||||
"x": -140,
|
||||
"y": 267,
|
||||
"width": 560,
|
||||
"height": 140,
|
||||
"color": "5"
|
||||
},
|
||||
{
|
||||
"id": "025be16f290f1830",
|
||||
"x": 501,
|
||||
"y": 280,
|
||||
"width": 520,
|
||||
"height": 293,
|
||||
"type": "file",
|
||||
"file": "2026-04-07 14-19-00.mkv"
|
||||
}
|
||||
],
|
||||
"edges": [
|
||||
{
|
||||
"id": "arrow-1",
|
||||
"fromNode": "zone-input",
|
||||
"fromSide": "right",
|
||||
"toNode": "zone-process",
|
||||
"toSide": "left",
|
||||
"color": "4",
|
||||
"label": "ingest"
|
||||
},
|
||||
{
|
||||
"id": "arrow-2",
|
||||
"fromNode": "zone-process",
|
||||
"fromSide": "right",
|
||||
"toNode": "zone-output",
|
||||
"toSide": "left",
|
||||
"color": "6",
|
||||
"label": "builds"
|
||||
}
|
||||
]
|
||||
}
|
||||
@@ -0,0 +1,251 @@
|
||||
{
|
||||
"nodes": [
|
||||
{
|
||||
"id": "zone-1", "type": "group", "label": "Beat 1 — Hook",
|
||||
"x": 0, "y": 0, "width": 1920, "height": 1080, "color": "1"
|
||||
},
|
||||
{
|
||||
"id": "z1-gif", "type": "file", "file": "claude-obsidian-gif-cover-16x9.gif",
|
||||
"x": 20, "y": 20, "width": 960, "height": 540
|
||||
},
|
||||
{
|
||||
"id": "z1-hook", "type": "text",
|
||||
"text": "# Your AI forgets everything.\n\nEvery new chat starts from zero.\n\nYou re-explain the same context.\nAgain. And again.\n\n> [!warning] The real cost\n> Tokens burned. Momentum lost.\n> Good answers buried in chat history.",
|
||||
"x": 1000, "y": 20, "width": 900, "height": 540, "color": "1"
|
||||
},
|
||||
{
|
||||
"id": "z1-bridge", "type": "text",
|
||||
"text": "## What if Claude just... knew?\n\nAnd got smarter every session?\n\nThat is what **claude-obsidian** does.",
|
||||
"x": 20, "y": 580, "width": 1880, "height": 480, "color": "3"
|
||||
},
|
||||
{
|
||||
"id": "z1-ref", "type": "text",
|
||||
"text": "## SCRIPT — Beat 1: Hook (0:00–0:35)\n\n**Opening line:** \"Your AI forgets everything. Every single chat starts from zero.\"\n\n**Show:** animated GIF playing. Let it loop 2-3x.\n\n**Say:**\n- \"You know the feeling. You explain your project, your stack, your goals. Every. Single. Time.\"\n- \"What if Claude just knew? And got smarter every session?\"\n- \"That is what claude-obsidian does. Let me show you.\"\n\n**Transition:** pan down to Beat 2",
|
||||
"x": 1960, "y": 0, "width": 600, "height": 1080, "color": "6"
|
||||
},
|
||||
|
||||
{
|
||||
"id": "zone-2", "type": "group", "label": "Beat 2 — How it works",
|
||||
"x": 0, "y": 1180, "width": 1920, "height": 1080, "color": "3"
|
||||
},
|
||||
{
|
||||
"id": "z2-title", "type": "text",
|
||||
"text": "# A wiki Claude writes for itself.",
|
||||
"x": 20, "y": 1200, "width": 1880, "height": 80, "color": "3"
|
||||
},
|
||||
{
|
||||
"id": "z2-img-a", "type": "file", "file": "_attachments/youtube/how-it-works.png",
|
||||
"x": 20, "y": 1300, "width": 940, "height": 528
|
||||
},
|
||||
{
|
||||
"id": "z2-img-b", "type": "file", "file": "_attachments/youtube/how-it-works-B.png",
|
||||
"x": 980, "y": 1300, "width": 920, "height": 528
|
||||
},
|
||||
{
|
||||
"id": "z2-flow-a", "type": "text",
|
||||
"text": "## 💬 You chat\n\nAsk questions.\nWork on projects.\nExplore ideas.",
|
||||
"x": 20, "y": 1848, "width": 580, "height": 192, "color": "5"
|
||||
},
|
||||
{
|
||||
"id": "z2-flow-b", "type": "text",
|
||||
"text": "## 📄 Claude files it\n\nPicks the note type.\nWrites wikilinks.\nUpdates the index.",
|
||||
"x": 620, "y": 1848, "width": 580, "height": 192, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "z2-flow-c", "type": "text",
|
||||
"text": "## 💬 Next session\n\nClaude reads the wiki first.\nZero re-explaining.\nSmarter from the start.",
|
||||
"x": 1220, "y": 1848, "width": 680, "height": 192, "color": "5"
|
||||
},
|
||||
{
|
||||
"id": "z2-layers", "type": "text",
|
||||
"text": "> [!tip] Key idea\n> **🔥 hot.md** — ~500 words, loaded every session automatically. Nearly free.\n> **🗂️ index.md** — full map of every page. Read when hot.md is not enough.\n> **📚 wiki/ pages** — the real knowledge. Only the relevant pages get read each session.",
|
||||
"x": 20, "y": 2060, "width": 1880, "height": 180
|
||||
},
|
||||
{
|
||||
"id": "z2-ref", "type": "text",
|
||||
"text": "## SCRIPT — Beat 2: How it works (0:35–1:30)\n\n**Opening line:** \"Here is how it works.\"\n\n**Show:** the three-layer diagram image on the right\n\n**Say:**\n- \"claude-obsidian turns Obsidian into a wiki that Claude writes for itself.\"\n- \"Every chat you have, the good stuff gets filed. Next session Claude reads it first.\"\n- \"Three layers: hot.md loads every session — about 500 words, nearly free. Then index.md for the full map. Then individual wiki pages, only the ones that matter.\"\n- \"You stay in budget. The wiki keeps growing.\"\n\n**Show:** open wiki/hot.md — show it is small and focused\n\n**Transition:** \"So how do you actually build this wiki? Three commands.\"\n\n**Files to show:** `wiki/hot.md`, `wiki/index.md`",
|
||||
"x": 1960, "y": 1180, "width": 600, "height": 1080, "color": "6"
|
||||
},
|
||||
|
||||
{
|
||||
"id": "zone-3", "type": "group", "label": "Beat 3 — /save",
|
||||
"x": 0, "y": 2360, "width": 1920, "height": 1080, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "z3-title", "type": "text",
|
||||
"text": "# /save — File the conversation into the wiki.",
|
||||
"x": 20, "y": 2380, "width": 1880, "height": 80, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "z3-img-a", "type": "file", "file": "_attachments/youtube/save-A.png",
|
||||
"x": 20, "y": 2480, "width": 940, "height": 530
|
||||
},
|
||||
{
|
||||
"id": "z3-img-b", "type": "file", "file": "_attachments/youtube/save-B.png",
|
||||
"x": 980, "y": 2480, "width": 920, "height": 530
|
||||
},
|
||||
{
|
||||
"id": "z3-bullets", "type": "text",
|
||||
"text": "## What it does\n\n- Reads the full current chat\n- Picks the right note type: concept, question, source, or decision\n- Writes frontmatter and wikilinks automatically\n- Files the note in the correct wiki folder\n- Updates index.md, the session log, and hot.md\n\n> [!tip] Result\n> Good answers stop disappearing.",
|
||||
"x": 20, "y": 3030, "width": 920, "height": 390
|
||||
},
|
||||
{
|
||||
"id": "z3-example", "type": "text",
|
||||
"text": "> [!example] Try it\n> `/save`\n>\n> or just: *save this as a concept note*\n\nClaude picks the type, writes the note,\nfiles it, and cross-links it. One command.",
|
||||
"x": 960, "y": 3030, "width": 940, "height": 390, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "z3-ref", "type": "text",
|
||||
"text": "## SCRIPT — Beat 3: /save (1:30–2:45)\n\n**Opening line:** \"Command one: /save.\"\n\n**Show:** type `/save` in the terminal while on a chat with good content\n\n**Say:**\n- \"One command. Claude reads the whole conversation.\"\n- \"It picks the right note type — concept, question, decision, source. Writes frontmatter. Adds wikilinks to related pages. Files it in the right folder.\"\n- \"It also updates the index and the hot cache, so next session it's already loaded.\"\n- \"Good answers stop disappearing.\"\n\n**Show:** the new note open in Obsidian — show frontmatter and wikilinks\n\n**Transition:** \"But what if you want Claude to go do the research for you?\"\n\n**Files to show:** fresh page in `wiki/concepts/`\n**Skill source:** `skills/save/SKILL.md`",
|
||||
"x": 1960, "y": 2360, "width": 600, "height": 1080, "color": "6"
|
||||
},
|
||||
|
||||
{
|
||||
"id": "zone-4", "type": "group", "label": "Beat 4 — /autoresearch",
|
||||
"x": 0, "y": 3540, "width": 1920, "height": 1080, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "z4-title", "type": "text",
|
||||
"text": "# /autoresearch — Autonomous research loop.",
|
||||
"x": 20, "y": 3560, "width": 1880, "height": 80, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "z4-img-a", "type": "file", "file": "_attachments/youtube/autoresearch-A.png",
|
||||
"x": 20, "y": 3660, "width": 940, "height": 530
|
||||
},
|
||||
{
|
||||
"id": "z4-img-b", "type": "file", "file": "_attachments/youtube/autoresearch-B.png",
|
||||
"x": 980, "y": 3660, "width": 920, "height": 530
|
||||
},
|
||||
{
|
||||
"id": "z4-bullets", "type": "text",
|
||||
"text": "## What it does\n\n- Breaks the topic into 3 to 5 research angles\n- Runs web searches and fetches the best sources\n- Extracts claims, entities, and concepts\n- Runs a second pass to fill in the gaps it noticed\n- Files a synthesis page + source pages + concept pages, all cross-linked\n\n> [!tip] Result\n> Research that becomes a wiki, not a tab graveyard.",
|
||||
"x": 20, "y": 4210, "width": 920, "height": 390
|
||||
},
|
||||
{
|
||||
"id": "z4-example", "type": "text",
|
||||
"text": "> [!example] Try it\n> `/autoresearch LLM wiki pattern`\n>\n> Output: 5 to 15 wiki pages, cross-linked, with citations.\n\nClaude works autonomously.\nYou get coffee.",
|
||||
"x": 960, "y": 4210, "width": 940, "height": 390, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "z4-ref", "type": "text",
|
||||
"text": "## SCRIPT — Beat 4: /autoresearch (2:45–4:15)\n\n**Opening line:** \"Command two: /autoresearch. This one feels like magic.\"\n\n**Show:** type `/autoresearch claude-obsidian best practices` — let it start running\n\n**Say:**\n- \"It breaks the topic into research angles. Searches the web. Fetches real sources. Pulls out the key claims.\"\n- \"Then it runs a second pass to fill in the gaps it noticed the first time.\"\n- \"End result: 5 to 15 wiki pages. Cross-linked. With citations. In about 3 minutes.\"\n- \"You do not get a chat bubble that disappears. You get a permanent knowledge base.\"\n\n**Show:** Obsidian sidebar filling up with new pages while it runs\n\n**Transition:** \"And the third command is what you are actually watching right now.\"\n\n**Files to show:** `wiki/sources/`, `wiki/concepts/`\n**Skill source:** `skills/autoresearch/SKILL.md`",
|
||||
"x": 1960, "y": 3540, "width": 600, "height": 1080, "color": "6"
|
||||
},
|
||||
|
||||
{
|
||||
"id": "zone-5", "type": "group", "label": "Beat 5 — /canvas",
|
||||
"x": 0, "y": 4720, "width": 1920, "height": 1080, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "z5-title", "type": "text",
|
||||
"text": "# /canvas — The visual layer of the wiki.",
|
||||
"x": 20, "y": 4740, "width": 1880, "height": 80, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "z5-img-a", "type": "file", "file": "_attachments/youtube/canvas-A.png",
|
||||
"x": 20, "y": 4840, "width": 940, "height": 530
|
||||
},
|
||||
{
|
||||
"id": "z5-img-b", "type": "file", "file": "_attachments/youtube/canvas-B.png",
|
||||
"x": 980, "y": 4840, "width": 920, "height": 530
|
||||
},
|
||||
{
|
||||
"id": "z5-bullets", "type": "text",
|
||||
"text": "## What it does\n\n- Creates infinite visual boards inside Obsidian\n- Drops in images, text cards, PDFs, and wiki notes\n- Positions nodes automatically inside named zones\n- Uses JSON Canvas 1.0 — open standard, portable and future-proof\n- Pairs with /banana for AI-generated images on any beat\n\n> [!key-insight] Meta moment\n> This canvas you are watching right now was built with /canvas.",
|
||||
"x": 20, "y": 5390, "width": 920, "height": 390
|
||||
},
|
||||
{
|
||||
"id": "z5-example", "type": "text",
|
||||
"text": "> [!example] Try it\n> `/canvas new youtube-video`\n> `/canvas add note LLM Wiki Pattern`\n> `/canvas add image _attachments/youtube/save-A.png`\n\nInstant visual board.\nObsidian-native. Shareable.",
|
||||
"x": 960, "y": 5390, "width": 940, "height": 390, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "z5-ref", "type": "text",
|
||||
"text": "## SCRIPT — Beat 5: /canvas (4:15–5:30)\n\n**Opening line:** \"Command three: /canvas. And what you are watching right now is it.\"\n\n**Show:** zoom out on this full canvas — let them see all 7 zones laid out\n\n**Say:**\n- \"This is /canvas. It gives the wiki a visual layer.\"\n- \"Drop in images, text cards, wiki notes, PDFs. Claude positions everything automatically.\"\n- \"You can build explainers, mood boards, project maps, video scripts — like this one.\"\n- \"This canvas? Every zone, every node — built with /canvas.\"\n\n**Show live:** type `/canvas add note LLM Wiki Pattern` — watch a node appear\n\n**Transition:** \"So that is the three commands. Let me quickly show you how the memory system behind them works.\"\n\n**Skill source:** `skills/canvas/SKILL.md`",
|
||||
"x": 1960, "y": 4720, "width": 600, "height": 1080, "color": "6"
|
||||
},
|
||||
|
||||
{
|
||||
"id": "zone-6", "type": "group", "label": "Beat 6 — Under the hood",
|
||||
"x": 0, "y": 5900, "width": 1920, "height": 1080, "color": "3"
|
||||
},
|
||||
{
|
||||
"id": "z6-title", "type": "text",
|
||||
"text": "# Under the hood — why token cost stays low as the wiki grows.",
|
||||
"x": 20, "y": 5920, "width": 1880, "height": 80, "color": "3"
|
||||
},
|
||||
{
|
||||
"id": "z6-graph", "type": "file", "file": "image-example-graph-view.png",
|
||||
"x": 20, "y": 6020, "width": 920, "height": 518
|
||||
},
|
||||
{
|
||||
"id": "z6-wikimap", "type": "file", "file": "image-example-wiki-map-view.png",
|
||||
"x": 960, "y": 6020, "width": 940, "height": 518
|
||||
},
|
||||
{
|
||||
"id": "z6-hot", "type": "text",
|
||||
"text": "## 🔥 hot.md\nLoaded automatically every session.\n~500 words. Near-zero token cost.\nRecent context. What is coming next.",
|
||||
"x": 20, "y": 6558, "width": 600, "height": 200, "color": "1"
|
||||
},
|
||||
{
|
||||
"id": "z6-index", "type": "text",
|
||||
"text": "## 🗂️ index.md\nOne-line summary of every wiki page.\nRead when hot.md is not enough.\nThe full map, compactly.",
|
||||
"x": 640, "y": 6558, "width": 620, "height": 200, "color": "3"
|
||||
},
|
||||
{
|
||||
"id": "z6-pages", "type": "text",
|
||||
"text": "## 📚 wiki/ pages\nConcepts. Sources. Decisions. People.\nOnly the relevant pages get read.\nDeep knowledge, on demand.",
|
||||
"x": 1280, "y": 6558, "width": 620, "height": 200, "color": "4"
|
||||
},
|
||||
{
|
||||
"id": "z6-insight", "type": "text",
|
||||
"text": "> [!key-insight] The wiki can grow to thousands of pages. Your token cost per session barely moves.",
|
||||
"x": 20, "y": 6778, "width": 1880, "height": 162
|
||||
},
|
||||
{
|
||||
"id": "z6-ref", "type": "text",
|
||||
"text": "## SCRIPT — Beat 6: Under the hood (5:30–6:30)\n\n**Opening line:** \"So how does Claude actually remember things without blowing your token budget?\"\n\n**Show:** pan across — graph view left, wiki map right, three layers below\n\n**Say:**\n- \"Three layers. That is the whole system.\"\n- \"hot.md loads every session — about 500 words. Recent context. Nearly free.\"\n- \"index.md is the full map: one line per page. Claude reads it when hot.md is not enough.\"\n- \"Then the actual wiki pages — only the ones that matter for that session.\"\n- \"The wiki can grow to thousands of pages. Your bill barely moves.\"\n\n**Show:** `wiki/hot.md` and `wiki/index.md` briefly\n\n**Transition:** pan down to install",
|
||||
"x": 1960, "y": 5900, "width": 600, "height": 1080, "color": "6"
|
||||
},
|
||||
|
||||
{
|
||||
"id": "zone-7", "type": "group", "label": "Beat 7 — Install + CTA",
|
||||
"x": 0, "y": 7080, "width": 1920, "height": 1080, "color": "5"
|
||||
},
|
||||
{
|
||||
"id": "z7-title", "type": "text",
|
||||
"text": "# Install in 2 lines — Free. Open source. Works with Claude Code today.",
|
||||
"x": 20, "y": 7100, "width": 1880, "height": 80, "color": "5"
|
||||
},
|
||||
{
|
||||
"id": "z7-gif", "type": "file", "file": "claude-obsidian-gif-1x1.gif",
|
||||
"x": 20, "y": 7200, "width": 480, "height": 480
|
||||
},
|
||||
{
|
||||
"id": "z7-cover", "type": "file", "file": "claude-obsidian-cover-16x9.png",
|
||||
"x": 520, "y": 7200, "width": 1380, "height": 480
|
||||
},
|
||||
{
|
||||
"id": "z7-cmd", "type": "text",
|
||||
"text": "```bash\nclaude plugin marketplace add AgriciDaniel/claude-obsidian\nclaude plugin install claude-obsidian@agricidaniel-claude-obsidian\n```",
|
||||
"x": 20, "y": 7700, "width": 920, "height": 160
|
||||
},
|
||||
{
|
||||
"id": "z7-links", "type": "text",
|
||||
"text": "## Links\n\n- **GitHub**: github.com/AgriciDaniel/claude-obsidian\n- **Community**: skool.com/ai-marketing-hub-pro\n- **Website**: agricidaniel.com\n\n> [!tip] Subscribe\n> More Claude Code builds every week.",
|
||||
"x": 960, "y": 7700, "width": 940, "height": 440, "color": "5"
|
||||
},
|
||||
{
|
||||
"id": "z7-ref", "type": "text",
|
||||
"text": "## SCRIPT — Beat 7: Install + CTA (6:30–8:00)\n\n**Opening line:** \"Installing takes literally two lines.\"\n\n**Show:** terminal — type both install commands live, slowly\n\n**Say:**\n- \"Copy these two lines. Paste. Done.\"\n- \"claude-obsidian is free, open source, and works with Claude Code right now.\"\n- \"GitHub link is in the description. Come join the Skool community — free tier, live sessions every week.\"\n- \"If this was useful, hit subscribe. I build with Claude Code every week.\"\n\n**End:** let the animated GIF loop while you close out.\n\n**Do NOT rush this beat.**",
|
||||
"x": 1960, "y": 7080, "width": 600, "height": 1080, "color": "6"
|
||||
}
|
||||
],
|
||||
"edges": [
|
||||
{"id": "e-flow-ab", "fromNode": "z2-flow-a", "fromSide": "right", "toNode": "z2-flow-b", "toSide": "left", "fromEnd": "none", "toEnd": "arrow"},
|
||||
{"id": "e-flow-bc", "fromNode": "z2-flow-b", "fromSide": "right", "toNode": "z2-flow-c", "toSide": "left", "fromEnd": "none", "toEnd": "arrow"},
|
||||
{"id": "e-hot-index", "fromNode": "z6-hot", "fromSide": "right", "toNode": "z6-index", "toSide": "left", "fromEnd": "none", "toEnd": "arrow"},
|
||||
{"id": "e-index-pages", "fromNode": "z6-index", "fromSide": "right", "toNode": "z6-pages", "toSide": "left", "fromEnd": "none", "toEnd": "arrow"}
|
||||
]
|
||||
}
|
||||
@@ -0,0 +1,145 @@
|
||||
---
|
||||
type: comparison
|
||||
title: "Claude + Obsidian Ecosystem: Feature Matrix"
|
||||
created: 2026-04-08
|
||||
updated: 2026-04-08
|
||||
tags:
|
||||
- ecosystem
|
||||
- competitive-analysis
|
||||
- claude-obsidian
|
||||
- cherry-picks
|
||||
status: current
|
||||
related:
|
||||
- "[[cherry-picks]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Andrej Karpathy]]"
|
||||
sources:
|
||||
- "[[claude-obsidian-ecosystem-research]]"
|
||||
---
|
||||
|
||||
# Claude + Obsidian Ecosystem: Feature Matrix
|
||||
|
||||
> Researched 2026-04-08 | 16+ projects analyzed | See [[cherry-picks]] for action items
|
||||
|
||||
---
|
||||
|
||||
## Legend
|
||||
- ✅ Has it
|
||||
- ❌ Missing
|
||||
- 🟡 Partial
|
||||
- ⭐ Best-in-class implementation
|
||||
|
||||
---
|
||||
|
||||
## LLM Wiki Pattern Projects (Claude Code Skills)
|
||||
|
||||
| Feature | claude-obsidian | claudesidian | llm-knowledge-bases | llm-wiki | obsidian-wiki | obsidian-claude-pkm |
|
||||
|---------|:-:|:-:|:-:|:-:|:-:|:-:|
|
||||
| /wiki setup & scaffold | ✅ | 🟡 `/init-bootstrap` | 🟡 `/kb-init` | ✅ | ✅ setup.sh | 🟡 `/onboard` |
|
||||
| Source ingestion | ✅ | ❌ | ✅ | ✅ | ✅ | ❌ |
|
||||
| Wiki query | ✅ | ❌ | ✅ 3 depths | 🟡 | ✅ | ❌ |
|
||||
| Wiki lint | ✅ | ❌ | ✅ | ✅ | ❌ | ❌ |
|
||||
| /save conversation | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ |
|
||||
| Autoresearch loop | ✅ | ❌ | 🟡 | ❌ | ❌ | ❌ |
|
||||
| Canvas / visual layer | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ |
|
||||
| Hot cache | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ |
|
||||
| **Delta tracking** | ❌ | ❌ | ❌ | ❌ | ✅⭐ | ❌ |
|
||||
| **Multi-depth query** | ❌ | ❌ | ✅⭐ | ❌ | ❌ | ❌ |
|
||||
| **URL ingestion** | ❌ | 🟡 firecrawl | ❌ | ✅ | ✅ | ❌ |
|
||||
| **Vision / image ingest** | ❌ | 🟡 gemini | ❌ | ❌ | ✅⭐ | ❌ |
|
||||
| **Auto-commit hooks** | ❌ | ❌ | ❌ | ✅ git | ❌ | ✅⭐ |
|
||||
| **Marp / slides output** | ❌ | ❌ | ✅⭐ | ✅ | ❌ | ❌ |
|
||||
| **Chart output** | ❌ | ❌ | ✅ matplotlib | ❌ | ❌ | ❌ |
|
||||
| **Hybrid search (BM25+vec)** | ❌ | ❌ | ❌ | ✅⭐ qmd | ❌ | ❌ |
|
||||
| **Goal cascade (PKM)** | ❌ | 🟡 PARA | ❌ | ❌ | ❌ | ✅⭐ |
|
||||
| **Daily/weekly review** | ❌ | 🟡 | ❌ | ❌ | ❌ | ✅⭐ |
|
||||
| **Adopt existing vault** | ❌ | ✅⭐ | ❌ | ❌ | ❌ | ✅⭐ |
|
||||
| **Multi-agent compat.** | ❌ | ❌ | ❌ | ❌ | ✅⭐ | ❌ |
|
||||
| **X/Twitter ingest** | ❌ | ❌ | ✅⭐ smaug | ❌ | ❌ | ❌ |
|
||||
| Marketplace install | ✅ | ❌ | ✅ | ❌ | ❌ | ❌ |
|
||||
| Public repo | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
|
||||
|
||||
---
|
||||
|
||||
## Native Obsidian Plugins (Embedded UI)
|
||||
|
||||
| Feature | Claudian | Nexus (claudesidian-mcp) | infio-copilot |
|
||||
|---------|:-:|:-:|:-:|
|
||||
| Sidebar chat in Obsidian | ✅ | ✅ | ✅ |
|
||||
| Inline edit w/ diff | ✅⭐ word-level | ✅ | ✅ |
|
||||
| Plan mode | ✅⭐ Shift+Tab | ❌ | ❌ |
|
||||
| @mention files/agents | ✅⭐ | ❌ | 🟡 |
|
||||
| MCP server support | ✅ | ✅⭐ external | ❌ |
|
||||
| Multi-tab conversations | ✅ | ❌ | ❌ |
|
||||
| Workspace memory | ❌ | ✅⭐ JSONL | ✅ workspaces |
|
||||
| Task management | ❌ | ✅⭐ | ❌ |
|
||||
| Semantic search | ❌ | ✅ | ✅⭐ local embed |
|
||||
| PDF → Markdown | ❌ | ✅ | ❌ |
|
||||
| Web page capture | ❌ | ✅ | ❌ |
|
||||
| Mobile support | ❌ | ✅⭐ | ❌ |
|
||||
| Obsidian Sync compatible | N/A | ✅⭐ | N/A |
|
||||
| Stars / popularity | ~200 est. | ~800 est. | ~300 est. |
|
||||
|
||||
---
|
||||
|
||||
## MCP Servers
|
||||
|
||||
| Server | Key Differentiator | Requires |
|
||||
|--------|-------------------|----------|
|
||||
| obsidian-mcp-tools | Templater execution + SLSA attestation | Local REST API + Smart Connections |
|
||||
| obsidian-memory-mcp | AI memories as Markdown in graph view | Node 18+ |
|
||||
| obsidian-claude-code-mcp | WebSocket, auto-discovers vaults | Claude Code |
|
||||
| administrativetrick/obsidian-mcp | Minimal, simple | Claude Desktop |
|
||||
| MarkusPfundstein/mcp-obsidian | Via REST API | Local REST API |
|
||||
|
||||
---
|
||||
|
||||
## kepano/obsidian-skills (Special — from Obsidian Creator)
|
||||
|
||||
Linus Kepano (Obsidian creator + Minimal theme) published official Agent Skills for Obsidian:
|
||||
|
||||
| Skill | What It Teaches |
|
||||
|-------|----------------|
|
||||
| obsidian-markdown | Obsidian Flavored Markdown (callouts, embeds, wikilinks, properties) |
|
||||
| obsidian-bases | Obsidian Bases (.base files, views, filters, formulas) |
|
||||
| json-canvas | JSON Canvas spec (.canvas nodes/edges/groups) |
|
||||
| obsidian-cli | Vault management via Obsidian CLI |
|
||||
| defuddle | Extract clean Markdown from web pages (saves tokens) |
|
||||
|
||||
> **Key signal**: This project validates that the Agent Skills format is the right standard.
|
||||
> These skills are platform-agnostic (Claude Code, Codex, OpenCode).
|
||||
|
||||
---
|
||||
|
||||
## Popularity Snapshot (Traditional Plugins)
|
||||
|
||||
| Plugin | Stars | Approach |
|
||||
|--------|-------|---------|
|
||||
| obsidian-copilot | 5,776 | Multi-provider vault chat |
|
||||
| obsidian-smart-connections | 4,357 | Semantic search + embeddings |
|
||||
| obsidian-textgenerator-plugin | 1,837 | Text generation |
|
||||
| chatgpt-md | 1,229 | Chat in Markdown |
|
||||
| obsidian-local-gpt | 569 | Local LLM |
|
||||
| obsidian-ai-tools | 272 | Supabase + OpenAI semantic search |
|
||||
|
||||
---
|
||||
|
||||
## Where claude-obsidian Wins
|
||||
|
||||
1. **Hot cache** — session context mechanism is unique in the ecosystem
|
||||
2. **Canvas skill** — no other LLM Wiki project has a visual layer
|
||||
3. **Marketplace install** — most polished install experience
|
||||
4. **/save conversation** — filing chat sessions as wiki pages is unique
|
||||
5. **Pub quality docs** — README, install guide, demo GIFs
|
||||
6. **Dual repo** (public + community) — distribution model unique
|
||||
|
||||
## Where claude-obsidian Has Gaps
|
||||
|
||||
See [[cherry-picks]] for prioritized list with implementation notes.
|
||||
|
||||
Top 5 gaps by impact:
|
||||
1. No delta tracking → re-ingests everything on every compile
|
||||
2. No URL ingestion → forces manual copy-paste
|
||||
3. No auto-commit → vault changes aren't versioned automatically
|
||||
4. No multi-depth query → one mode for all questions
|
||||
5. No vision input → can't ingest screenshots or images
|
||||
@@ -0,0 +1,69 @@
|
||||
---
|
||||
type: concept
|
||||
title: "Compounding Knowledge"
|
||||
complexity: basic
|
||||
domain: knowledge-management
|
||||
aliases:
|
||||
- "Knowledge Compounding"
|
||||
- "Persistent Synthesis"
|
||||
created: 2026-04-07
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- concept
|
||||
- knowledge-management
|
||||
status: mature
|
||||
related:
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Hot Cache]]"
|
||||
- "[[Andrej Karpathy]]"
|
||||
- "[[concepts/_index]]"
|
||||
sources:
|
||||
---
|
||||
|
||||
# Compounding Knowledge
|
||||
|
||||
The central insight behind the [[LLM Wiki Pattern]]: knowledge in a wiki compounds like interest in a bank. Every source added, every question answered, every analysis filed makes the wiki more valuable — not just by adding pages, but by enriching the connections between existing pages.
|
||||
|
||||
---
|
||||
|
||||
## Why Normal AI Chats Don't Compound
|
||||
|
||||
In a standard chat, knowledge is ephemeral. Each session starts fresh. Even if you upload the same documents repeatedly, the LLM re-derives the same insights from scratch. Nothing accumulates.
|
||||
|
||||
The same is true of most RAG systems: they index raw documents and retrieve chunks at query time. The retrieval gets the right fragments, but no synthesis is built up. Nothing is compiled. Ask the same complex question twice and you get the same assembly process twice.
|
||||
|
||||
---
|
||||
|
||||
## How Wiki Knowledge Compounds
|
||||
|
||||
When a new source arrives, the LLM doesn't just index it. It integrates it:
|
||||
- Updates entity pages with new information
|
||||
- Flags contradictions with existing claims
|
||||
- Strengthens or challenges the evolving synthesis
|
||||
- Adds cross-references from the new source to existing pages and back
|
||||
|
||||
The cross-references are already there next time. The contradictions have already been flagged. The synthesis already reflects everything that was read.
|
||||
|
||||
**The wiki is pre-compiled knowledge.** RAG re-compiles on every query.
|
||||
|
||||
---
|
||||
|
||||
## The Maintenance Problem
|
||||
|
||||
Wikis maintained by humans decay. The maintenance burden grows faster than the value — updating cross-references, keeping summaries current, noting when new data contradicts old claims. Humans abandon wikis because no one wants to do the bookkeeping.
|
||||
|
||||
LLMs don't get bored. They don't forget to update a cross-reference. The cost of maintenance is near zero. This is the practical reason the wiki pattern works: the entity that's best at the tedious maintenance work is the same entity that reads and writes the wiki.
|
||||
|
||||
---
|
||||
|
||||
## In Practice
|
||||
|
||||
One X user turned 383 scattered files and over 100 meeting transcripts into a compact wiki and dropped token usage by 95% when querying with Claude. The drop came from two sources: better navigation (index + hot cache vs. full document search) and pre-compiled synthesis (no re-deriving the same insights from scratch).
|
||||
|
||||
---
|
||||
|
||||
## Connections
|
||||
|
||||
See [[LLM Wiki Pattern]] for the full architecture.
|
||||
See [[Hot Cache]] for the session context mechanism.
|
||||
See [[Andrej Karpathy]] for the origin of this framing.
|
||||
@@ -0,0 +1,295 @@
|
||||
---
|
||||
type: concept
|
||||
title: "DragonScale Memory"
|
||||
address: c-000001
|
||||
complexity: advanced
|
||||
domain: knowledge-management
|
||||
aliases:
|
||||
- "DragonScale"
|
||||
- "DragonScale Architecture"
|
||||
- "Fractal Memory"
|
||||
created: 2026-04-23
|
||||
updated: 2026-04-24
|
||||
tags:
|
||||
- concept
|
||||
- knowledge-management
|
||||
- memory
|
||||
- architecture
|
||||
- fractal
|
||||
status: proposed
|
||||
related:
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Compounding Knowledge]]"
|
||||
- "[[Hot Cache]]"
|
||||
- "[[concepts/_index]]"
|
||||
sources:
|
||||
---
|
||||
|
||||
# DragonScale Memory
|
||||
|
||||
A memory-layer design for LLM wiki vaults, inspired by the Heighway dragon curve. Four mechanisms (fold operator, deterministic page addresses, semantic tiling, boundary-first autoresearch) give an LLM-maintained wiki a principled way to grow, compact, and stay coherent. The dragon curve is a design-justification device, not a reasoning architecture.
|
||||
|
||||
> **Status: v0.4 2026-04-24.** All four mechanisms shipped as opt-in features. Phase 0 (spec) + Phase 1 (wiki-fold skill, dry-run verified) + Phase 2 (address MVP) + Phase 3 (semantic tiling) + Phase 3.5/3.6 (hardening) + Phase 4 (boundary-first autoresearch). See Review History for the progression.
|
||||
|
||||
---
|
||||
|
||||
## Scope
|
||||
|
||||
DragonScale is a **memory architecture**: it governs how a wiki grows, compacts, addresses its pages, and checks for duplicates. It is **not a search, planning, or reasoning algorithm.** Agent reasoning uses existing patterns (Tree of Thoughts with BFS/DFS/beam search; Yao et al. 2023).
|
||||
|
||||
**Honest disclaimer**: memory-layer choices are never neutral with respect to reasoning. What the vault surfaces, and in what order, shapes what the model sees. Long-context performance is position-sensitive (Liu et al. 2023, *Lost in the Middle*), and MemGPT's premise is that paging policy affects task success (Packer et al. 2023). One of the four mechanisms below (boundary-first autoresearch) explicitly crosses into agenda control; it is included deliberately and marked as such.
|
||||
|
||||
---
|
||||
|
||||
## The Core Analogy
|
||||
|
||||
Four dragon-curve properties map onto memory-system patterns already validated in adjacent fields. The word is *analogue*, not *identity*.
|
||||
|
||||
| Dragon curve property | Memory analogue | Strength of analogy |
|
||||
|---|---|---|
|
||||
| Paper-folding recursion: `D_{n+1} = D_n · R · swap(reverse(D_n))` | Hierarchical rollup / materialized summary with exponential fanout | Loose. Shares exponential batch structure, not compaction semantics. |
|
||||
| Turn derivable from bits of `n` (regular paperfolding sequence, OEIS A014577) | Deterministic page addresses as organizational convention (MVP is a creation-order counter, not a true content hash) | Loose. Deterministic addressing is useful independent of the dragon. |
|
||||
| Tiling / no self-intersection | Canonical-home coverage: one concept, one page | Medium. Dedup lint enforces this mechanically. |
|
||||
| Boundary dim ≈ 1.523627 vs interior dim 2 | Agent attention weighted toward frontier pages | Aesthetic. The fractal dimension number does no load-bearing work. |
|
||||
|
||||
The curve is useful for deciding *which knobs to tighten and why*, not as a math proof that any given mechanism is optimal.
|
||||
|
||||
---
|
||||
|
||||
## Mechanism 1 — Fold Operator
|
||||
|
||||
After a batch of ingests, run a fold: produce a meta-page summarizing the batch, link children back, update the index. Folds stack: after enough level-`k` folds accumulate, a level-`k+1` fold produces a super-summary.
|
||||
|
||||
This is a **hierarchical rollup**, loosely similar to LSM-tree compaction but with important differences.
|
||||
|
||||
**What it shares with LSM compaction:**
|
||||
- Exponential batch fanout across levels (like LevelDB's fixed level-size ratio, typically 10× per level in leveled mode)
|
||||
- Periodic consolidation rather than per-write work
|
||||
|
||||
**What it does NOT inherit from LSM:**
|
||||
- No sorted-key semantics (pages have semantic, not key-ordered, identity)
|
||||
- No SSTable/memtable distinction, no tombstones, no Bloom filters
|
||||
- No write-amplification arithmetic; no read-path acceleration
|
||||
- **Folds are additive**: children remain in place. LSM compaction rewrites and deletes. A DragonScale fold is closer to a materialized view than a compaction.
|
||||
|
||||
**Trigger options:**
|
||||
- `2^k` entry count (k=4 ⇒ every 16 log entries). Simple to implement; straightforward level math; ignores page size and novelty.
|
||||
- **Adaptive trigger (preferred for production)**: token budget (e.g., fold when unfolded batch exceeds N tokens), novelty score (average embedding distance from existing summaries), or staleness age (last fold > T days). Phase 1 will implement entry-count for MVP; adaptive triggers are a follow-up.
|
||||
|
||||
**Invariants:**
|
||||
- Idempotent on the same range (re-running is a no-op).
|
||||
- Reversible (children stay; a fold is additive).
|
||||
- Level-bounded: with entry-count trigger `2^k`, fold depth is at most `⌈log₂(N)⌉` above leaf pages. Derived, not empirical.
|
||||
|
||||
---
|
||||
|
||||
## Mechanism 2 — Deterministic Page Addresses
|
||||
|
||||
Every new page gets a stable `address` field in frontmatter. The Phase 2 MVP uses a simple creation-order counter:
|
||||
|
||||
```yaml
|
||||
address: c-000042
|
||||
```
|
||||
|
||||
Format: `c-<6-digit-counter>`. `c-` means "creation-order counter." Zero-padded.
|
||||
|
||||
**Future extension** (documented, not shipped in Phase 2):
|
||||
- Fold-relative path: `f1.2/c-000042` once folds exist, where `f1.2` encodes the fold-tree lineage.
|
||||
- Content hash suffix: `c-000042:h7f3c2` once the hash-rotation policy is decided.
|
||||
|
||||
**What Phase 2 MVP gives:**
|
||||
- Uniqueness: counter is monotonically increasing; deleted pages' addresses are retired, never reused.
|
||||
- Stability: never changes across content edits.
|
||||
- Determinism: derivable from the counter state at `.vault-meta/address-counter.txt`.
|
||||
- Ordering: preserves creation sequence.
|
||||
|
||||
**What this does NOT give (renamed "content-addressable paths" was misleading in v0.1):**
|
||||
- **No content-addressability in the MVP.** The Phase 2 address is a sequence counter, not a content hash. Renaming this mechanism from "content-addressable paths" to "deterministic page addresses" is more honest about what actually ships.
|
||||
- **No prompt cache benefit** (already corrected in v0.1 → v0.2). Per Anthropic docs, cache hits require byte-identical prefixes; an address field in frontmatter only helps if the frontmatter itself is inside a cached block AND stays byte-identical. Stable prefixes, not addresses, drive cache hits.
|
||||
|
||||
**Phase 2 exclusions** (all deferred):
|
||||
- Backfill of legacy pre-Phase-2 pages (will use `l-` prefix with its own counter).
|
||||
- Fold-ancestry bit prefix (requires committed folds from a future fold-of-folds skill).
|
||||
- Content hash suffix (rotation policy unresolved; see limitations).
|
||||
|
||||
**Implementation** (Phase 2, shipped):
|
||||
- `scripts/allocate-address.sh`: flock-guarded atomic allocator. All counter reads/writes go through this script; direct Write/Edit on `.vault-meta/address-counter.txt` is prohibited (would fire PostToolUse hook).
|
||||
- `skills/wiki-ingest/SKILL.md` → Address Assignment section: opt-in feature detection; delegates allocation to the helper; records path-to-address mapping in `.raw/.manifest.json` `address_map` for re-ingest stability.
|
||||
- `skills/wiki-lint/SKILL.md` → Address Validation section: format check, uniqueness check, counter-drift check, address-map consistency check.
|
||||
|
||||
**Lint severity model** (matches `skills/wiki-lint/SKILL.md` Address Validation behavior):
|
||||
- Post-rollout pages (frontmatter `created:` >= 2026-04-23, or any page newly created after DragonScale adoption) that lack an address are **errors**. This is the silent-regression guard.
|
||||
- Legacy pages (`created:` < 2026-04-23) without addresses are **informational**. The optional `.vault-meta/legacy-pages.txt` manifest can grandfather pages whose `created:` metadata is wrong or missing.
|
||||
- Meta pages (`_index.md`, `index.md`, `log.md`, `hot.md`, etc.) and fold pages are excluded entirely.
|
||||
|
||||
---
|
||||
|
||||
## Mechanism 3 — Semantic Tiling Lint
|
||||
|
||||
The tiling property says the same concept should live in one canonical page. Enforce it with an embedding-based dedup check in `wiki-lint`.
|
||||
|
||||
**Procedure (calibrated, not a guess):**
|
||||
1. Compute embeddings for every page. Default model: local `nomic-embed-text` via ollama on `http://127.0.0.1:11434`. Cost: local hardware time only (no API fees). The script supports a remote override under `--allow-remote-ollama`; remote endpoints may incur provider API fees.
|
||||
2. Compute pairwise cosine similarities for all page pairs.
|
||||
3. **Calibration** (one-time, before first use): label 50-100 in-vault page pairs as duplicate/near/distinct; find the thresholds that optimize target precision for each band.
|
||||
4. **Default bands** (used before calibration, then refined):
|
||||
- `≥ 0.90` — near-duplicate, lint error
|
||||
- `0.80 – 0.90` — review bucket, lint warning
|
||||
- `< 0.80` — distinct, no flag
|
||||
5. Never auto-merge. Output a review list.
|
||||
|
||||
**Why not a fixed 0.85?** v0.1 used 0.85 with no justification. Published thresholds in the embeddings literature span a wide range (Sentence Transformers' `community_detection` defaults to 0.75; Quora-duplicate calibrations land around 0.77–0.83; sparse-model defaults differ again). Thresholds are model-, corpus-, and objective-dependent, so calibration is required.
|
||||
|
||||
---
|
||||
|
||||
## Mechanism 4 — Boundary-First Autoresearch
|
||||
|
||||
> **Status: shipped (Phase 4, opt-in)** as of 2026-04-24. Implementation: `scripts/boundary-score.py`. Integration: `skills/autoresearch/SKILL.md` Topic Selection section B. Tests: `tests/test_boundary_score.py`.
|
||||
|
||||
Boundary pages (high out-degree relative to in-degree, recency-weighted) are the vault's frontier. `/autoresearch` invoked without a topic reads the top-5 boundary pages and offers them as research candidates; the user selects one (or types a free-text topic, or declines all and falls back to the original ask-user mode).
|
||||
|
||||
**Formula (exact)**:
|
||||
|
||||
```
|
||||
out_degree(p) = count of distinct filename-stem wikilinks in body of p that resolve to scoreable pages
|
||||
in_degree(p) = count of distinct scoreable pages whose body contains a wikilink to p
|
||||
recency_weight(p) = exp(-days_since_updated / 30) # no floor; old pages approach 0
|
||||
boundary_score(p) = (out_degree - in_degree) * recency_weight
|
||||
```
|
||||
|
||||
**Link resolution**: filename-stem only. `[[Foo]]` resolves to `Foo.md` anywhere in the vault. Aliases declared via frontmatter `aliases:` are NOT parsed. Folder-qualified links (e.g. `[[notes/Foo]]`) are resolved by stem alone. This matches Obsidian's default behavior for unique filenames but does not implement full alias resolution.
|
||||
|
||||
**Scoreable** = any page NOT excluded by any of:
|
||||
- frontmatter `type: meta` or `type: fold`
|
||||
- filename in `{_index.md, index.md, log.md, hot.md, overview.md, dashboard.md, Wiki Map.md, getting-started.md}`
|
||||
- path prefix in `wiki/folds/` or `wiki/meta/`
|
||||
- symlinks or paths whose resolved target escapes the vault root (rejected at scan time)
|
||||
|
||||
**Code-block filtering**: triple-backtick AND triple-tilde fenced code blocks are skipped, with CommonMark-like length tracking so a longer opening fence is not closed by a shorter inner fence. Indented code blocks (4+ spaces) are NOT filtered because Obsidian bullet lists commonly use 4-space indentation and contain real wikilinks. See `scripts/boundary-score.py:RECENCY_HALFLIFE_DAYS` for the sole tunable constant.
|
||||
|
||||
**Honest labeling**: this mechanism is **agenda control**, not pure memory. It shapes what the agent researches next. It is included in DragonScale because it is a direct consequence of the dragon-curve boundary analogy, and because it pairs naturally with folds (freshly folded pages have low out-degree; frontier pages are pre-fold). But the "memory only, not reasoning" framing does not cover it. Users who want a strict memory-layer subset should omit this mechanism (simply do not invoke `/autoresearch` without a topic, or do not set up `scripts/boundary-score.py`).
|
||||
|
||||
**What is NOT included**:
|
||||
- No auto-triggering. `/autoresearch` is still user-invoked.
|
||||
- No persistent boundary-score cache. Scoring is O(N * avg_links) and runs on every invocation from fresh wiki/ state.
|
||||
- No integration with folds or addresses. Pure graph analysis on the wikilink graph.
|
||||
- No automatic topic selection without user confirmation. The helper presents choices; the user picks.
|
||||
|
||||
---
|
||||
|
||||
## Operational Policies (required before implementation)
|
||||
|
||||
Adversarial review flagged these gaps in v0.1. Each must be decided before the corresponding phase ships.
|
||||
|
||||
| Policy | Phase 0 position | Decision point |
|
||||
|---|---|---|
|
||||
| **Retention / GC** | No automatic deletion. Pages are permanent. | Revisit if vault exceeds ~5000 pages. |
|
||||
| **Tombstones** | None. Deleted pages are removed via git revert. | Revisit if delete events become common. |
|
||||
| **Versioning** | Relied on git history, not in-vault versioning. | Address-hash rotation policy doubles as a coarse version signal. |
|
||||
| **Conflict resolution for contradictory folds** | Meta-page must quote both sources with explicit "conflict" callout. No automatic resolution. | Phase 1 spec required. |
|
||||
| **Concurrency / atomicity** | Single-writer assumption (one Claude session at a time). PostToolUse auto-commit serializes. | Multi-writer case deferred. |
|
||||
| **Provenance for meta-pages** | Every fold page must include frontmatter listing children and fold level. | Phase 1 must enforce. |
|
||||
| **Access control** | Out of scope. This is a single-user vault. | Revisit only if shared. |
|
||||
|
||||
---
|
||||
|
||||
## Mapping to Claude-Obsidian
|
||||
|
||||
| Mechanism | Status | New | Extends |
|
||||
|---|---|---|---|
|
||||
| Fold operator | shipped (Phase 1, dry-run verified) | `skills/wiki-fold/` | reads `log.md`, writes `wiki/folds/`, updates `index.md` on commit |
|
||||
| Address anchors | shipped (Phase 2, opt-in) | `scripts/allocate-address.sh`, new frontmatter field | `wiki-ingest` (assignment), `wiki-lint` (validation) |
|
||||
| Semantic tiling | shipped (Phase 2/3, opt-in) | `scripts/tiling-check.py`, `.vault-meta/tiling-thresholds.json` | `wiki-lint` with banded thresholds, calibration procedure documented |
|
||||
| Boundary-first | shipped (Phase 4, opt-in) | `scripts/boundary-score.py`, `tests/test_boundary_score.py` | `skills/autoresearch/SKILL.md` Topic Selection section B; `commands/autoresearch.md` no-topic path |
|
||||
|
||||
The existing hot → index → domain → page hierarchy already implements self-similarity across scales. That's the one dragon-curve property this vault had before DragonScale.
|
||||
|
||||
---
|
||||
|
||||
## Why This Over Alternatives
|
||||
|
||||
| Pattern | What it gives | What DragonScale adds |
|
||||
|---|---|---|
|
||||
| MemGPT virtual context (two-tier paging) | Main context ↔ external context swap | More than two levels; explicit fold triggers; dedup lint |
|
||||
| Pure LSM compaction | Exponential write-path throughput | Semantic-layer mechanisms (tiling, boundary); additive rollups over destructive merges |
|
||||
| Ad-hoc `/save` | Human-triggered filing | Rule-based fold cadence |
|
||||
| Vector-only RAG | Retrieval | Canonical-home structure; lineage addresses |
|
||||
|
||||
DragonScale composes patterns validated in adjacent systems: LSM *batching* (databases), MemGPT *paging* (agents), Anthropic *cache ordering* (prompt engineering), and embedding *dedup* (knowledge graphs).
|
||||
|
||||
---
|
||||
|
||||
## Known Limitations (v0.3)
|
||||
|
||||
- **Unvalidated at scale.** All four mechanisms are theoretical; none tested on a multi-thousand-page vault.
|
||||
- **Fold cadence is a knob, not a theorem.** `k=4` is a starting guess. Adaptive triggers are likely better.
|
||||
- **Address stability is unsolved.** Hash rotation on edits is a known issue; deferred.
|
||||
- **Boundary-first crosses scope.** Included with a warning, not quietly.
|
||||
- **Calibration load.** Tiling requires a one-time labeling pass; without it, only defaults apply.
|
||||
|
||||
---
|
||||
|
||||
## Primary Sources
|
||||
|
||||
Verified against primary sources on 2026-04-23. **Scope of tagging**: the specific numeric values, formulas, and named patterns below are tagged **[sourced]** when directly citable, **[derived]** when derivable from sourced material, or **[conjecture]** when based on reasoning without a specific source. **Not tagged** (and readers should treat as interpretive synthesis): framing sentences in the body such as "composes patterns validated," "self-similarity already exists," and the design rationale tying the four mechanisms together. These are editorial, not source-backed.
|
||||
|
||||
**Dragon curve math [sourced]**
|
||||
- Boundary dimension `2·log₂(λ)` where `λ³ − λ² − 2 = 0`, giving 1.523627086: [Dragon curve, Wikipedia](https://en.wikipedia.org/wiki/Dragon_curve)
|
||||
- Paper-folding construction and OEIS A014577: [Regular paperfolding sequence, Wikipedia](https://en.wikipedia.org/wiki/Regular_paperfolding_sequence); [OEIS A014577](https://oeis.org/A014577)
|
||||
- Tiling and rep-tiles: [Wolfram Demonstrations: Tiling Dragons and Rep-tiles of Order Two](https://demonstrations.wolfram.com/TilingDragonsAndRepTilesOfOrderTwo/)
|
||||
|
||||
**LSM trees [sourced]**
|
||||
- Level size ratios and compaction semantics: [RocksDB Compaction wiki](https://github.com/facebook/rocksdb/wiki/Compaction), [RocksDB Tuning Guide](https://github.com/facebook/rocksdb/wiki/RocksDB-Tuning-Guide), [How to Grow an LSM-tree? (2025)](https://arxiv.org/abs/2504.17178)
|
||||
- LevelDB 10× level ratio: referenced in the arXiv paper above. Treat as *typical*, not required.
|
||||
|
||||
**LLM memory architectures [sourced]**
|
||||
- OS-inspired paging: [MemGPT: Towards LLMs as Operating Systems (Packer et al. 2023)](https://arxiv.org/abs/2310.08560)
|
||||
- Position sensitivity: [Lost in the Middle (Liu et al. 2023)](https://direct.mit.edu/tacl/article/doi/10.1162/tacl_a_00638/119630/Lost-in-the-Middle-How-Language-Models-Use-Long)
|
||||
- Note-based agentic memory: [A-Mem (2025)](https://arxiv.org/abs/2502.12110)
|
||||
|
||||
**Prompt caching [sourced]**
|
||||
- Byte-identical prefix requirement, breakpoint mechanics, TTL options: [Anthropic Prompt Caching docs](https://platform.claude.com/docs/en/build-with-claude/prompt-caching)
|
||||
|
||||
**Embedding thresholds [sourced]**
|
||||
- Sentence Transformers defaults and calibration examples: [Sentence Transformers util](https://sbert.net/docs/package_reference/util.html), [SBERT evaluation docs](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html)
|
||||
|
||||
**Reasoning search (out of scope, cited only to justify the scope boundary) [sourced]**
|
||||
- [Tree of Thoughts (Yao et al. 2023)](https://arxiv.org/abs/2305.10601)
|
||||
|
||||
**Items marked [conjecture] in this doc:**
|
||||
- `k=4`/`k=5` starting value for fold cadence (needs empirical tuning)
|
||||
- `~30s` full-vault embedding-pass time (needs measurement)
|
||||
- `boundary_score` formula exact weighting (a plausible starting form; not validated against retrieval metrics)
|
||||
|
||||
**Items marked [derived]:**
|
||||
- `⌈log₂(N)⌉` fold-depth bound (trivially derivable from the entry-count trigger)
|
||||
- Default tiling bands `{≥0.90, 0.80-0.90, <0.80}` before calibration (interpolated from cited ranges in Sentence Transformers examples; not optimal by construction)
|
||||
|
||||
---
|
||||
|
||||
## Review History
|
||||
|
||||
**v0.1 (2026-04-23, initial draft)** — written after a verification pass against Wikipedia, arXiv, and Anthropic docs. Four mechanisms proposed.
|
||||
|
||||
**v0.4 (2026-04-24, Phase 4 shipped)** — Mechanism 4 (boundary-first autoresearch) implemented as `scripts/boundary-score.py` with `tests/test_boundary_score.py` covering parsing, recency weight, wikilink extraction (with fence-length + tilde + indented-block tests), graph construction (self-loop/unresolved/meta-target exclusion), symlink rejection, and CLI surface (`--top`, `--page`, `--json`). Integrated into `skills/autoresearch/SKILL.md` as an opt-in Topic Selection mode with explicit helper-failure fallback. Spec's "NOT IMPLEMENTED" marker removed; exact scoring formula (no recency floor), filename-stem-only resolution disclosure, scope, and "what is NOT included" section added. Phase 3.6 pre-Phase-4 hardening shipped concurrently (5 fixes: `--report` path confinement, rollout baseline, AGENTS.md consistency, wiki-ingest .raw contradiction, install-guide version).
|
||||
|
||||
**v0.3 (2026-04-23, Phase 2 alignment)** — Mechanism 2 rewritten to match the actual Phase 2 MVP shipped in `wiki-ingest` and `wiki-lint`. Renamed from "Content-Addressable Paths" to "Deterministic Page Addresses" (the MVP is a creation-order counter, not a content hash). Documented the extension path for fold-ancestry bits and content-hash suffix, both explicitly deferred.
|
||||
|
||||
**v0.2 (2026-04-23, post-adversarial review)** — after `codex exec` adversarial review. All 7 critiques accepted:
|
||||
|
||||
1. *LSM "structurally identical"* → weakened to "loosely analogous to hierarchical rollup"; non-inherited properties listed explicitly.
|
||||
2. *Prompt cache address benefit* → removed strong claim; narrowed to organizational convention.
|
||||
3. *0.85 threshold* → replaced with calibration procedure and banded defaults.
|
||||
4. *2^k cadence* → justified as implementation convenience; adaptive trigger flagged as preferred for production.
|
||||
5. *Scope boundary contradiction* → acknowledged; boundary-first explicitly labeled as agenda control.
|
||||
6. *Missing production mechanisms* → added Operational Policies section (retention, versioning, conflict resolution, concurrency, provenance).
|
||||
7. *Unverified claims* → tagged specific numeric values, formulas, and named patterns as [sourced], [derived], or [conjecture]. Editorial synthesis in the body explicitly flagged as not tagged (see scope note under Primary Sources).
|
||||
|
||||
---
|
||||
|
||||
## Connections
|
||||
|
||||
See [[LLM Wiki Pattern]] for the broader pattern this extends.
|
||||
See [[Compounding Knowledge]] for why persistent state is the precondition for DragonScale.
|
||||
See [[Hot Cache]] for the existing 500-word session context, which is a level-0 manual fold.
|
||||
See [[Andrej Karpathy]] for the intellectual lineage.
|
||||
@@ -0,0 +1,95 @@
|
||||
---
|
||||
type: concept
|
||||
title: "Hot Cache"
|
||||
complexity: basic
|
||||
domain: knowledge-management
|
||||
aliases:
|
||||
- "hot.md"
|
||||
- "Session Cache"
|
||||
- "Context Cache"
|
||||
created: 2026-04-07
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- concept
|
||||
- knowledge-management
|
||||
- context
|
||||
status: mature
|
||||
related:
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Compounding Knowledge]]"
|
||||
- "[[index]]"
|
||||
- "[[hot]]"
|
||||
- "[[concepts/_index]]"
|
||||
sources:
|
||||
---
|
||||
|
||||
# Hot Cache
|
||||
|
||||
A ~500-word summary of the most recent context in the wiki vault. Stored in `wiki/hot.md`. Updated at the end of every session and after every significant ingest or query.
|
||||
|
||||
The hot cache exists to answer one question: "where did we leave off?" A new session reads `hot.md` first. If the answer is there, it skips crawling the rest of the wiki.
|
||||
|
||||
---
|
||||
|
||||
## What It Stores
|
||||
|
||||
- What was most recently ingested or discussed
|
||||
- Key recent facts and takeaways
|
||||
- Pages recently created or updated
|
||||
- Active threads and open questions
|
||||
- What the user is currently focused on
|
||||
|
||||
---
|
||||
|
||||
## Format
|
||||
|
||||
```markdown
|
||||
---
|
||||
type: meta
|
||||
title: "Hot Cache"
|
||||
updated: YYYY-MM-DDTHH:MM:SS
|
||||
---
|
||||
|
||||
# Recent Context
|
||||
|
||||
## Last Updated
|
||||
YYYY-MM-DD — [what happened]
|
||||
|
||||
## Key Recent Facts
|
||||
- [Most important recent takeaway]
|
||||
- [Second]
|
||||
|
||||
## Recent Changes
|
||||
- Created: new wiki pages from this ingest
|
||||
- Updated: existing pages with new connections
|
||||
- Flagged: contradictions between sources where found
|
||||
|
||||
## Active Threads
|
||||
- User is researching [topic]
|
||||
- Open question: [thing being investigated]
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Rules
|
||||
|
||||
- Keep it under 500 words. It is a cache, not a journal.
|
||||
- Overwrite it completely each time. Not append-only.
|
||||
- One file. Not split by date.
|
||||
- Updated after every ingest, significant query, and at the end of each session.
|
||||
|
||||
---
|
||||
|
||||
## Why It Matters
|
||||
|
||||
Without the hot cache, every session starts cold: read the index (1000 tokens), read several domain sub-indexes, read several individual pages. With the hot cache, the first 500 tokens often have everything needed.
|
||||
|
||||
In practice, adding `hot.md` to an executive assistant vault dramatically reduces the token cost of session startup compared to crawling multiple wiki pages.
|
||||
|
||||
The hot cache is especially valuable in cross-project setups: another Claude Code project can point at this vault and read `hot.md` first to get recent context at minimal token cost.
|
||||
|
||||
---
|
||||
|
||||
## Connections
|
||||
|
||||
The hot cache is part of the [[LLM Wiki Pattern]] token discipline strategy. See [[index]] for how the broader navigation works.
|
||||
@@ -0,0 +1,97 @@
|
||||
---
|
||||
type: concept
|
||||
title: "LLM Wiki Pattern"
|
||||
complexity: intermediate
|
||||
domain: knowledge-management
|
||||
aliases:
|
||||
- "LLM Knowledge Base"
|
||||
- "Karpathy Wiki"
|
||||
- "Persistent Wiki"
|
||||
created: 2026-04-07
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- concept
|
||||
- knowledge-management
|
||||
- llm
|
||||
- obsidian
|
||||
status: mature
|
||||
related:
|
||||
- "[[Hot Cache]]"
|
||||
- "[[Compounding Knowledge]]"
|
||||
- "[[Andrej Karpathy]]"
|
||||
- "[[index]]"
|
||||
- "[[concepts/_index]]"
|
||||
sources:
|
||||
---
|
||||
|
||||
# LLM Wiki Pattern
|
||||
|
||||
A pattern for building persistent, compounding knowledge bases using LLMs. Originated by [[Andrej Karpathy]]. The key insight: instead of re-deriving knowledge from raw documents on every query (RAG), the LLM incrementally builds and maintains a structured wiki that gets richer with every source added.
|
||||
|
||||
---
|
||||
|
||||
## The Core Idea
|
||||
|
||||
Most AI knowledge tools work like RAG: index raw documents, retrieve chunks at query time, generate an answer. Nothing accumulates. Ask a question that needs five documents and the LLM reassembles fragments every time.
|
||||
|
||||
The wiki pattern is different. When a new source arrives, the LLM reads it, extracts what matters, and integrates it into the wiki: updating entity pages, noting contradictions, strengthening the synthesis. The cross-references are already there. The knowledge is compiled once and kept current.
|
||||
|
||||
**The wiki is a persistent, compounding artifact.** The human curates sources and asks questions. The LLM writes and maintains everything.
|
||||
|
||||
---
|
||||
|
||||
## Three Layers
|
||||
|
||||
```
|
||||
.raw/ Layer 1 — immutable source documents
|
||||
wiki/ Layer 2 — LLM-generated knowledge base
|
||||
CLAUDE.md Layer 3 — schema that tells the LLM how to maintain it
|
||||
```
|
||||
|
||||
The LLM owns Layer 2 entirely. It creates pages, updates them when new sources arrive, maintains cross-references, and keeps everything consistent. The human reads; the LLM writes.
|
||||
|
||||
---
|
||||
|
||||
## Operations
|
||||
|
||||
**Ingest** — drop a source into `.raw/`, tell the LLM to process it. The LLM reads the source, discusses key takeaways, writes a summary page, updates entity and concept pages, and logs the operation. One source typically touches 8-15 wiki pages.
|
||||
|
||||
**Query** — ask a question. The LLM reads the index to find relevant pages, synthesizes an answer with citations. Good answers get filed back into the wiki.
|
||||
|
||||
**Lint** — periodic health check. Find orphan pages, dead links, stale claims, missing cross-references.
|
||||
|
||||
---
|
||||
|
||||
## Index and Log
|
||||
|
||||
**index.md** — content-oriented. A catalog of all pages with one-line summaries, organized by category. The LLM reads this first on every query to find relevant pages.
|
||||
|
||||
**log.md** — chronological. Append-only record of every ingest, query, and lint pass. Parseable: `grep "^## \[" log.md | head -10`
|
||||
|
||||
---
|
||||
|
||||
## Why It Works
|
||||
|
||||
The tedious part of maintaining a knowledge base is bookkeeping: updating cross-references, noting when new data contradicts old claims, keeping summaries current. Humans abandon wikis because the maintenance burden grows faster than the value. LLMs don't get bored. The wiki stays maintained because the cost of maintenance is near zero.
|
||||
|
||||
At small scale (~100 sources, ~hundreds of pages), the index file is sufficient. No vector database, no embeddings, no infrastructure. Just markdown files.
|
||||
|
||||
---
|
||||
|
||||
## Comparison to RAG
|
||||
|
||||
| Dimension | LLM Wiki | Semantic RAG |
|
||||
|-----------|----------|-------------|
|
||||
| Finding | Reads index, follows links | Similarity search over embeddings |
|
||||
| Infrastructure | Just markdown files | Embedding model + vector DB |
|
||||
| Cost | Tokens only | Ongoing compute + storage |
|
||||
| Maintenance | Run a lint | Re-embed when content changes |
|
||||
| Scale limit | Hundreds of pages | Millions of documents |
|
||||
|
||||
---
|
||||
|
||||
## Connections
|
||||
|
||||
See [[Compounding Knowledge]] for why the pattern produces more value over time.
|
||||
See [[Hot Cache]] for the session context optimization.
|
||||
See [[Andrej Karpathy]] for the pattern's origin.
|
||||
@@ -0,0 +1,44 @@
|
||||
---
|
||||
type: concept
|
||||
title: "Persistent Wiki Artifact"
|
||||
created: 2026-04-24
|
||||
updated: 2026-04-24
|
||||
tags:
|
||||
- llm-wiki
|
||||
- knowledge-management
|
||||
- agent-memory
|
||||
status: developing
|
||||
related:
|
||||
- "[[How does the LLM Wiki pattern work?]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Compounding Knowledge]]"
|
||||
- "[[Source-First Synthesis]]"
|
||||
- "[[Query-Time Retrieval]]"
|
||||
---
|
||||
|
||||
# Persistent Wiki Artifact
|
||||
|
||||
A persistent wiki artifact is the maintained Markdown layer between raw sources and future questions. In Karpathy's LLM Wiki description, the agent reads source material, extracts key information, and integrates it into an interlinked wiki instead of only retrieving chunks at answer time: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
|
||||
## Boundary Filled
|
||||
|
||||
The selected question explains that an LLM Wiki compounds knowledge, but it does not isolate the artifact as the unit of memory. This page makes that boundary explicit: memory is stored in files that can be browsed, linked, reviewed, and revised.
|
||||
|
||||
## Extracted Claims
|
||||
|
||||
- The LLM Wiki pattern defines raw sources, the generated wiki, and a schema document as separate layers: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
- In that pattern, the raw source collection is treated as immutable, while the wiki layer is owned and maintained by the LLM: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
- The pattern frames the wiki as a compounding artifact whose cross-references, contradiction flags, and synthesis persist across later questions: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
- Obsidian supports Wikilinks such as `[[Three laws of motion]]`, which lets Markdown files form an internal network of notes: https://obsidian.md/help/links
|
||||
- Obsidian can automatically update internal links when a file is renamed, depending on the vault setting: https://obsidian.md/help/links
|
||||
|
||||
## Implications for This Vault
|
||||
|
||||
- The durable memory object is the page, not the chat turn.
|
||||
- The page needs frontmatter, stable title, wikilinks, and source URLs so later agents can inspect provenance.
|
||||
- The page should remain small enough to revise directly, because the LLM Wiki pattern depends on updating existing synthesis when new sources arrive: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
|
||||
## Primary Sources
|
||||
|
||||
- https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
- https://obsidian.md/help/links
|
||||
@@ -0,0 +1,60 @@
|
||||
---
|
||||
type: concept
|
||||
title: "Pro Hub Challenge"
|
||||
created: 2026-04-14
|
||||
updated: 2026-04-14
|
||||
tags:
|
||||
- concept
|
||||
- community
|
||||
- ai-marketing-hub
|
||||
- claude-seo
|
||||
- open-source
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[Claude SEO]]"
|
||||
- "[[2026-04-14-claude-seo-v190-session]]"
|
||||
- "[[Semantic Topic Clustering]]"
|
||||
- "[[Search Experience Optimization]]"
|
||||
---
|
||||
|
||||
# Pro Hub Challenge
|
||||
|
||||
A community challenge hosted in the [AI Marketing Hub Pro](https://www.skool.com/ai-marketing-hub-pro) Skool community where members build extensions for Claude SEO or Claude Blog, competing for $600 in Claude Credits.
|
||||
|
||||
## First Challenge (v1.9.0, April 2026)
|
||||
|
||||
**6 submissions, 5 scored Proficient or above**
|
||||
|
||||
| Contributor | Submission | Score | Integrated? |
|
||||
|------------|------------|-------|-------------|
|
||||
| Lutfiya Miller | Semantic Cluster Engine | Winner | Yes — `seo-cluster` |
|
||||
| Florian Schmitz | SXO Skill | Proficient | Yes — `seo-sxo` |
|
||||
| Dan Colta | SEO Drift Monitor | Proficient | Yes — `seo-drift` |
|
||||
| Chris Muller | Multi-lingual Blog | Proficient | Partial — SEO parts into `seo-hreflang` |
|
||||
| Matej Marjanovic | E-commerce + Cost Config | Proficient | Yes — `seo-ecommerce` + cost guardrails |
|
||||
| Benjamin Samar | SEO Dungeon | Reviewed | No — not integrated in v1.9.0 |
|
||||
|
||||
## Integration Pattern
|
||||
|
||||
Community submissions go through:
|
||||
1. **Full code review** — architecture, quality, security
|
||||
2. **Security audit** — SSRF, injection, credential handling
|
||||
3. **Cherry-pick** — only SEO-relevant parts for claude-seo, blog parts stay for claude-blog
|
||||
4. **De-brand** — remove contributor-specific branding (e.g., ScienceExperts.ai)
|
||||
5. **Attribution** — `original_author` in SKILL.md frontmatter, HTML comments in agents, CONTRIBUTORS.md
|
||||
|
||||
## Submission Guidelines (from CONTRIBUTING.md)
|
||||
|
||||
1. SKILL.md under 500 lines, references under 200 lines
|
||||
2. All scripts must import `validate_url()` for SSRF protection
|
||||
3. Include `original_author` in SKILL.md frontmatter metadata
|
||||
4. Submit via PR or post in AI Marketing Hub community
|
||||
|
||||
## Second Challenge (April 2026)
|
||||
|
||||
**Keyword**: LEADS
|
||||
**Prize pool**: $600 ($400 first place, $200 second place) in Claude Credits
|
||||
**Deadline**: April 28, 2026
|
||||
**Scope**: Anything touching lead generation — Claude Code skills, n8n workflows, MCP servers, scrapers, dashboards, pipelines. If it helps someone capture, qualify, nurture, or convert leads, it counts.
|
||||
**Rules**: GitHub repo or .zip file + 1-2 minute demo video. Must be functional (not a concept). Solo or team both welcome.
|
||||
**Previous winner**: Lutfiya Miller (seo-cluster, integrated in v1.9.0)
|
||||
@@ -0,0 +1,44 @@
|
||||
---
|
||||
type: concept
|
||||
title: "Query-Time Retrieval"
|
||||
created: 2026-04-24
|
||||
updated: 2026-04-24
|
||||
tags:
|
||||
- rag
|
||||
- retrieval
|
||||
- llm-wiki
|
||||
status: developing
|
||||
related:
|
||||
- "[[How does the LLM Wiki pattern work?]]"
|
||||
- "[[Wiki vs RAG]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Persistent Wiki Artifact]]"
|
||||
- "[[Source-First Synthesis]]"
|
||||
---
|
||||
|
||||
# Query-Time Retrieval
|
||||
|
||||
Query-time retrieval is the baseline memory pattern that LLM Wiki is contrasted against: relevant material is retrieved when the user asks a question, and the answer is generated from the retrieved context.
|
||||
|
||||
## Boundary Filled
|
||||
|
||||
The selected question contrasts wiki accumulation with RAG, but it does not define the retrieval side precisely. This page anchors the contrast in the original RAG paper and in the LLM Wiki gist.
|
||||
|
||||
## Extracted Claims
|
||||
|
||||
- The RAG paper defines retrieval-augmented generation as combining parametric memory with non-parametric memory for language generation: https://arxiv.org/abs/2005.11401
|
||||
- The RAG paper describes the non-parametric memory as a dense vector index of Wikipedia accessed with a neural retriever: https://arxiv.org/abs/2005.11401
|
||||
- The paper reports that RAG models generated more specific, diverse, and factual language than a parametric-only seq2seq baseline in its evaluated generation tasks: https://arxiv.org/abs/2005.11401
|
||||
- Karpathy's LLM Wiki gist describes common document workflows as uploading files, retrieving relevant chunks at query time, and generating an answer: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
- Karpathy's LLM Wiki gist states that this query-time pattern makes the model rediscover and assemble knowledge on each question instead of accumulating synthesis: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
- The MemGPT paper frames limited LLM context windows as a constraint for extended conversations and document analysis, then proposes virtual context management across memory tiers: https://arxiv.org/abs/2310.08560
|
||||
|
||||
## Contrast With Wiki Memory
|
||||
|
||||
Query-time retrieval can provide external evidence at answer time. The LLM Wiki pattern shifts part of the work earlier by compiling source material into maintained pages before later queries arrive: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
|
||||
## Primary Sources
|
||||
|
||||
- https://arxiv.org/abs/2005.11401
|
||||
- https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
- https://arxiv.org/abs/2310.08560
|
||||
@@ -0,0 +1,43 @@
|
||||
---
|
||||
type: concept
|
||||
title: "SEO Drift Monitoring"
|
||||
created: 2026-04-14
|
||||
updated: 2026-04-14
|
||||
tags:
|
||||
- concept
|
||||
- seo
|
||||
- monitoring
|
||||
- change-detection
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[Claude SEO]]"
|
||||
- "[[Pro Hub Challenge]]"
|
||||
---
|
||||
|
||||
# SEO Drift Monitoring
|
||||
|
||||
"Git for SEO" — captures baselines of SEO-critical page elements, then diffs against current state to detect regressions. Contributed to [[Claude SEO]] v1.9.0 by Dan Colta.
|
||||
|
||||
## What It Tracks
|
||||
|
||||
17 comparison rules across 3 severity levels:
|
||||
|
||||
| Severity | Examples |
|
||||
|----------|----------|
|
||||
| CRITICAL | Schema removed, canonical changed, noindex added, H1 removed |
|
||||
| WARNING | Title changed, CWV regression >20%, meta description changed |
|
||||
| INFO | H2 structure changed, content hash changed, image count changed |
|
||||
|
||||
## Architecture
|
||||
|
||||
- **SQLite persistence** at `~/.cache/claude-seo/drift/baselines.db`
|
||||
- **4 Python scripts**: `drift_baseline.py` (capture), `drift_compare.py` (diff), `drift_report.py` (HTML report), `drift_history.py` (timeline)
|
||||
- **Security-hardened**: uses only `fetch_page.py` for URL fetching (SSRF-protected). Original submission had a curl fallback that bypassed SSRF protection — completely removed during integration.
|
||||
|
||||
## Commands
|
||||
|
||||
```
|
||||
/seo drift baseline <url> # Capture current state
|
||||
/seo drift compare <url> # Compare against baseline
|
||||
/seo drift history <url> # Show all checks over time
|
||||
```
|
||||
@@ -0,0 +1,159 @@
|
||||
---
|
||||
type: concept
|
||||
title: "SVG Diagram Style Guide"
|
||||
created: 2026-04-14
|
||||
updated: 2026-04-14
|
||||
tags:
|
||||
- design
|
||||
- svg
|
||||
- brand
|
||||
- diagrams
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[index]]"
|
||||
sources:
|
||||
- "claude-ads/assets/diagrams/ (17 SVGs, v1.5.0)"
|
||||
---
|
||||
|
||||
# SVG Diagram Style Guide
|
||||
|
||||
The canonical visual style for all diagrams across agricidaniel's Claude Code skill repos. Extracted from the 17 production SVGs in claude-ads. Use this as the reference when creating or updating diagrams for any skill repo.
|
||||
|
||||
## Font
|
||||
|
||||
```
|
||||
font-family: 'Space Grotesk', system-ui, -apple-system, sans-serif
|
||||
```
|
||||
|
||||
Space Grotesk is the only typeface. No fallback to serif or monospace.
|
||||
|
||||
## Color Palette
|
||||
|
||||
### Core (use these in every diagram)
|
||||
|
||||
| Token | Hex | Role |
|
||||
|-------|-----|------|
|
||||
| bg | #0A0A0A | Canvas background (near-black) |
|
||||
| card | #111111 | Card/container fill |
|
||||
| card-inner | #1A1A1A | Nested element fill |
|
||||
| border | #2D2D2D | Card borders, dividers |
|
||||
| text-primary | #F5F5F0 | Headings, labels (off-white) |
|
||||
| text-secondary | #888888 | Descriptions, captions |
|
||||
| text-tertiary | #6a6a6a | De-emphasized metadata |
|
||||
| accent | #E07850 | Primary accent, arrows, highlights (warm rust-orange) |
|
||||
| accent-bright | #FF6B35 | Secondary accent, hover states (brighter orange) |
|
||||
|
||||
### Platform/Category Colors (use for variety within a diagram)
|
||||
|
||||
| Token | Hex | Typical use |
|
||||
|-------|-----|-------------|
|
||||
| blue | #60A5FA | Google, data, information |
|
||||
| purple | #8b5cf6 | Meta, strategy, creative |
|
||||
| cyan | #06b6d4 | LinkedIn, networking |
|
||||
| green | #4ADE80 | Success, validation, TikTok |
|
||||
| rose | #F43F5E | YouTube, alerts |
|
||||
| orange | #FF6B35 | Microsoft, secondary accent |
|
||||
| gray | #888888 | Neutral, generic platforms |
|
||||
|
||||
### Status Colors (for pass/warn/fail indicators)
|
||||
|
||||
| Token | Hex | Role |
|
||||
|-------|-----|------|
|
||||
| pass | #16a34a | Pass, success |
|
||||
| warn | #f59e0b | Warning, attention |
|
||||
| fail | #dc2626 | Fail, critical |
|
||||
|
||||
## Typography Scale
|
||||
|
||||
| Element | Size | Weight | Color | Extra |
|
||||
|---------|------|--------|-------|-------|
|
||||
| Diagram title | 16-17px | 700 | #F5F5F0 | text-anchor: middle |
|
||||
| Subtitle | 11px | 400 | #888888 | text-anchor: middle |
|
||||
| Section label | 13px | 700 | accent color | letter-spacing: 2 |
|
||||
| Card heading | 12-15px | 600-700 | #F5F5F0 | text-anchor: middle |
|
||||
| Card subtext | 9-11px | 400 | accent color | Skill/agent name |
|
||||
| Body text | 10px | 400 | #888888 | Descriptions |
|
||||
| Tiny label | 9px | 400 | #6a6a6a | Metadata, counts |
|
||||
|
||||
## Layout Primitives
|
||||
|
||||
### Outer Container
|
||||
```xml
|
||||
<rect width="800" height="500" fill="#0A0A0A"/>
|
||||
```
|
||||
Standard canvas is 800x500. Some diagrams use 900x250 or 900x350 depending on content.
|
||||
|
||||
### Card
|
||||
```xml
|
||||
<rect x="40" y="20" width="720" height="120" rx="16" fill="#111111" stroke="#2D2D2D" stroke-width="1.5"/>
|
||||
```
|
||||
- Corner radius: `rx="16"` for outer containers
|
||||
- Border: `#2D2D2D`, `stroke-width="1.5"`
|
||||
|
||||
### Colored Top Bar (card accent)
|
||||
```xml
|
||||
<rect x="40" y="20" width="720" height="4" rx="2" fill="#E07850"/>
|
||||
```
|
||||
4px height, sits at the top edge of the card. Color indicates category.
|
||||
|
||||
### Inner Card (nested element)
|
||||
```xml
|
||||
<rect x="60" y="230" width="105" height="60" rx="6" fill="#1A1A1A" stroke="#2D2D2D" stroke-width="1"/>
|
||||
```
|
||||
- Corner radius: `rx="6"` for small inner cards, `rx="9"` for medium
|
||||
- Fill: `#1A1A1A` (slightly lighter than parent card)
|
||||
|
||||
### Numbered Circle (for sequences)
|
||||
```xml
|
||||
<circle cx="138" cy="60" r="14" fill="#0A0A0A" stroke="#60A5FA" stroke-width="1.5"/>
|
||||
<text x="138" y="60" font-size="12" fill="#60A5FA" text-anchor="middle" font-weight="bold" dominant-baseline="central">1</text>
|
||||
```
|
||||
Circle stroke color matches the step's category color.
|
||||
|
||||
### Arrow Connector
|
||||
```xml
|
||||
<line x1="400" y1="140" x2="400" y2="170" stroke="#E07850" stroke-width="1.5"/>
|
||||
<polygon points="394,167 400,177 406,167" fill="#E07850"/>
|
||||
```
|
||||
Always `#E07850`. Vertical for flow-down, horizontal for left-to-right pipelines.
|
||||
|
||||
### Horizontal Divider (title underline)
|
||||
```xml
|
||||
<line x1="380" y1="36" x2="520" y2="36" stroke="#E07850" stroke-width="2.5" stroke-linecap="round"/>
|
||||
```
|
||||
Short centered line under diagram title. Always accent color.
|
||||
|
||||
## Diagram Types (from claude-ads)
|
||||
|
||||
| # | Name | Layout | Size |
|
||||
|---|------|--------|------|
|
||||
| 01 | Architecture | 3-layer vertical stack | 800x500 |
|
||||
| 02 | Parallel Audit | Agent grid with flow | 800x500 |
|
||||
| 04 | Platform Checks | Checklist columns | 800x500 |
|
||||
| 05 | Quality Gates | Rule cards | 800x500 |
|
||||
| 06 | How It Works | Step sequence | 900x250 |
|
||||
| 07 | Data Flow | Horizontal pipeline | 900x250 |
|
||||
| 08 | Industry Templates | Card grid | 900x350 |
|
||||
| 10 | MCP Integration | Connection diagram | 800x500 |
|
||||
| 12 | Privacy Flow | Vertical flow | 800x500 |
|
||||
| 13 | Scoring Algorithm | Formula breakdown | 800x500 |
|
||||
| 14 | Creative Pipeline | 5-step horizontal | 900x250 |
|
||||
| 15 | Platform Grid | 2-row card grid | 900x350 |
|
||||
| 16 | PDF Pipeline | Process flow | 900x250 |
|
||||
| 17 | A/B Testing | Split comparison | 800x500 |
|
||||
| 18 | PPC Calculators | Tool cards | 900x350 |
|
||||
| 19 | Audit Lifecycle | Circular flow | 800x500 |
|
||||
| 20 | Install Methods | Option cards | 900x250 |
|
||||
|
||||
## Rules
|
||||
|
||||
1. Always dark theme. Never white or light backgrounds.
|
||||
2. Space Grotesk only. No other fonts.
|
||||
3. #E07850 is the signature accent. Use it for arrows, highlights, and the primary visual element.
|
||||
4. Cards always have #2D2D2D borders. Never borderless cards.
|
||||
5. Colored top bars (4px) identify categories. One color per category, consistent across the diagram.
|
||||
6. Text is always left-aligned or center-aligned. Never right-aligned.
|
||||
7. No gradients, shadows, or blur filters. Flat design only.
|
||||
8. Numbered circles for sequential steps. Color matches category.
|
||||
9. Arrow connectors are always #E07850 with triangle tips.
|
||||
10. File naming: zero-padded number prefix (01-, 02-, etc.) + kebab-case description.
|
||||
@@ -0,0 +1,44 @@
|
||||
---
|
||||
type: concept
|
||||
title: "Search Experience Optimization (SXO)"
|
||||
created: 2026-04-14
|
||||
updated: 2026-04-14
|
||||
tags:
|
||||
- concept
|
||||
- seo
|
||||
- ux
|
||||
- serp-analysis
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[Claude SEO]]"
|
||||
- "[[Pro Hub Challenge]]"
|
||||
- "[[Semantic Topic Clustering]]"
|
||||
---
|
||||
|
||||
# Search Experience Optimization (SXO)
|
||||
|
||||
A methodology that reads SERPs backwards to detect page-type mismatches, derives user stories from search features, and scores pages from persona perspectives. Contributed to [[Claude SEO]] v1.9.0 by Florian Schmitz.
|
||||
|
||||
## Core Insight
|
||||
|
||||
> "Read SERPs backwards" — instead of optimizing content FOR the SERP, analyze WHAT the SERP tells you about user expectations, then check if your page meets them.
|
||||
|
||||
## Process
|
||||
|
||||
1. **Page-type detection** — classify the URL as one of 8 types (Landing, Blog, Product, Hybrid, Service, Comparison, Local, Tool)
|
||||
2. **SERP pattern matching** — compare what Google shows (featured snippets, PAA, ads, related searches) against what the page provides
|
||||
3. **Mismatch detection** — if SERP says "users want comparison" but page is "product page", that's a mismatch
|
||||
4. **User story derivation** — from SERP features, derive 4-7 personas with emotional states, barriers, goals
|
||||
5. **Persona scoring** — score the page from each persona's perspective (0-100 across 4 dimensions)
|
||||
6. **Wireframe generation** — IST (current) vs SOLL (ideal) wireframes with ultra-concrete placeholders
|
||||
|
||||
## Key Innovation
|
||||
|
||||
Most SEO tools analyze pages in isolation. SXO uses the SERP as a proxy for user intent — the SERP IS the research that Google already did about what users want. This makes the analysis data-driven without needing user testing.
|
||||
|
||||
## Command
|
||||
|
||||
```
|
||||
/seo sxo <url>
|
||||
/seo sxo wireframe <url>
|
||||
```
|
||||
@@ -0,0 +1,46 @@
|
||||
---
|
||||
type: concept
|
||||
title: "Semantic Topic Clustering"
|
||||
created: 2026-04-14
|
||||
updated: 2026-04-14
|
||||
tags:
|
||||
- concept
|
||||
- seo
|
||||
- content-strategy
|
||||
- clustering
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[Claude SEO]]"
|
||||
- "[[Pro Hub Challenge]]"
|
||||
- "[[Search Experience Optimization]]"
|
||||
---
|
||||
|
||||
# Semantic Topic Clustering
|
||||
|
||||
SERP-based keyword grouping that replaces paid tools ($50-200/month) with Claude's reasoning. Contributed to [[Claude SEO]] v1.9.0 by Lutfiya Miller (Pro Hub Challenge Winner).
|
||||
|
||||
## How It Works
|
||||
|
||||
1. **Seed keyword** provided by user
|
||||
2. **SERP fetching** — get Google results for the seed and related terms (via WebSearch or DataForSEO)
|
||||
3. **Overlap scoring** — compare top-10 results between keyword pairs:
|
||||
- 7-10 overlapping URLs = same post (keyword cannibalization)
|
||||
- 4-6 overlapping = same cluster (supporting content)
|
||||
- 2-3 overlapping = interlink opportunity
|
||||
- 0-1 overlapping = separate clusters
|
||||
4. **Hub-spoke architecture** — 1 pillar page (2500-4000 words) + 2-5 clusters + 2-4 posts each
|
||||
5. **Internal link matrix** — bidirectional linking plan with backward link injection
|
||||
6. **Visualization** — interactive cluster-map.html (SVG, dark mode, keyboard accessible)
|
||||
|
||||
## Key Design Decisions
|
||||
|
||||
- **No Python scripts** — clustering is prompt-driven (Claude's reasoning + WebSearch)
|
||||
- **Optional execution** — outputs content briefs when claude-blog isn't installed, full pipeline when it is
|
||||
- **Resume capability** — for long multi-post execution runs
|
||||
- **DataForSEO integration** — uses `serp_organic_live_advanced` for live SERP data when available (with cost check)
|
||||
|
||||
## Command
|
||||
|
||||
```
|
||||
/seo cluster <seed-keyword>
|
||||
```
|
||||
@@ -0,0 +1,42 @@
|
||||
---
|
||||
type: concept
|
||||
title: "Source-First Synthesis"
|
||||
created: 2026-04-24
|
||||
updated: 2026-04-24
|
||||
tags:
|
||||
- llm-wiki
|
||||
- synthesis
|
||||
- provenance
|
||||
status: developing
|
||||
related:
|
||||
- "[[How does the LLM Wiki pattern work?]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Compounding Knowledge]]"
|
||||
- "[[Persistent Wiki Artifact]]"
|
||||
- "[[Query-Time Retrieval]]"
|
||||
---
|
||||
|
||||
# Source-First Synthesis
|
||||
|
||||
Source-first synthesis is the LLM Wiki practice of keeping raw sources separate from the generated wiki while requiring the wiki to cite and integrate those sources. Karpathy's pattern describes raw sources as the source of truth and the generated wiki as the maintained synthesis layer: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
|
||||
## Boundary Filled
|
||||
|
||||
The selected question says the wiki pattern integrates sources, but it does not spell out the provenance discipline. This page records the rule: synthesis is allowed to be rewritten, but source material remains the cited anchor.
|
||||
|
||||
## Extracted Claims
|
||||
|
||||
- Karpathy's LLM Wiki pattern says raw sources can include articles, papers, images, and data files, and that the LLM reads them without modifying them: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
- The same source describes the wiki as summaries, entity pages, concept pages, comparisons, overview, and synthesis maintained by the LLM: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
- The ingest operation can create a source summary, update indexes, update relevant entity and concept pages, and append a log entry: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
- The query operation reads relevant wiki pages and synthesizes answers with citations, and useful answers can be filed back into the wiki: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
- The RAG paper identifies provenance and updating world knowledge as open problems for knowledge-intensive generation systems: https://arxiv.org/abs/2005.11401
|
||||
|
||||
## Operating Rule
|
||||
|
||||
Source-first synthesis is stricter than unsourced summarization. A new concept page should identify the sources it used, state what was extracted from them, and avoid treating the generated page as a replacement for the source document.
|
||||
|
||||
## Primary Sources
|
||||
|
||||
- https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
|
||||
- https://arxiv.org/abs/2005.11401
|
||||
@@ -0,0 +1,43 @@
|
||||
---
|
||||
type: meta
|
||||
title: "Concepts Index"
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- meta
|
||||
- index
|
||||
- concept
|
||||
domain: knowledge-management
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[index]]"
|
||||
- "[[dashboard]]"
|
||||
- "[[Wiki Map]]"
|
||||
- "[[Hot Cache]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Compounding Knowledge]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Hot Cache]]"
|
||||
- "[[Compounding Knowledge]]"
|
||||
---
|
||||
|
||||
# Concepts Index
|
||||
|
||||
Navigation: [[index]] | [[entities/_index|Entities]] | [[sources/_index|Sources]]
|
||||
|
||||
All concept pages — ideas, patterns, and frameworks extracted from sources.
|
||||
|
||||
---
|
||||
|
||||
## Knowledge Management
|
||||
|
||||
- [[LLM Wiki Pattern]] — the core architecture for persistent, compounding knowledge bases
|
||||
- [[Hot Cache]] — ~500-word session context file, updated after every ingest
|
||||
- [[Compounding Knowledge]] — why the wiki grows more valuable over time, unlike RAG
|
||||
- [[DragonScale Memory]] — memory-layer spec: fold operator, deterministic page addresses, semantic tiling, boundary-first autoresearch (status: shipped v0.4, all four mechanisms opt-in)
|
||||
- [[Persistent Wiki Artifact]]: durable Markdown page as the LLM's memory object (developing)
|
||||
- [[Source-First Synthesis]]: provenance discipline for LLM wiki layers (developing)
|
||||
- [[Query-Time Retrieval]]: query synthesis with citations, complementary to Obsidian search (developing)
|
||||
|
||||
---
|
||||
|
||||
## Add new concepts here as they are extracted from sources.
|
||||
@@ -0,0 +1,156 @@
|
||||
---
|
||||
type: concept
|
||||
title: "Cherry-Picks: Feature Backlog from Ecosystem Research"
|
||||
created: 2026-04-08
|
||||
updated: 2026-04-08
|
||||
tags:
|
||||
- backlog
|
||||
- cherry-picks
|
||||
- product-roadmap
|
||||
- claude-obsidian
|
||||
status: current
|
||||
related:
|
||||
- "[[claude-obsidian-ecosystem]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
sources:
|
||||
- "[[claude-obsidian-ecosystem-research]]"
|
||||
---
|
||||
|
||||
# Cherry-Picks: Feature Backlog
|
||||
|
||||
> Sourced from ecosystem research 2026-04-08 | 16+ projects analyzed
|
||||
> Prioritized by: impact × implementation ease × uniqueness
|
||||
|
||||
---
|
||||
|
||||
## Tier 1 — Quick Wins (High Impact, Low Effort)
|
||||
|
||||
### 1. URL Ingestion in /wiki-ingest
|
||||
**Source**: ekadetov/llm-wiki, Ar9av/obsidian-wiki
|
||||
**What it is**: Pass a URL directly to ingest instead of a file path. Agent fetches the page, cleans it, saves to `.raw/`, then ingests.
|
||||
**Current state**: Users must manually copy-paste web content.
|
||||
**How to add**: Detect `https://` prefix in ingest skill → WebFetch → save to `.raw/articles/` → proceed with normal ingest.
|
||||
**Bonus**: Pair with **defuddle** (kepano's web cleaner) for clean token-efficient extraction.
|
||||
|
||||
### 2. Auto-Commit PostToolUse Hook
|
||||
**Source**: ballred/obsidian-claude-pkm, ekadetov/llm-wiki
|
||||
**What it is**: Every Write/Edit tool call in the vault triggers `git add -A && git commit -m "auto: [filename] [timestamp]"`.
|
||||
**Current state**: No auto-commit. Users must manually push.
|
||||
**How to add**: PostToolUse hook in hooks.json targeting Write + Edit tools, scoped to wiki/ directory.
|
||||
**Note**: Makes vault a proper version-controlled knowledge base automatically.
|
||||
|
||||
### 3. defuddle Web Cleaning Skill
|
||||
**Source**: kepano/obsidian-skills
|
||||
**What it is**: A skill that wraps `defuddle-cli` — strips ads, nav, clutter from web pages before ingest. Reduces token usage ~40-60% on typical web articles.
|
||||
**How to add**: New `defuddle` sub-skill or reference in wiki-ingest. Requires `defuddle-cli` npm package.
|
||||
|
||||
---
|
||||
|
||||
## Tier 2 — Medium Effort, High Value
|
||||
|
||||
### 4. Delta Tracking Manifest
|
||||
**Source**: Ar9av/obsidian-wiki
|
||||
**What it is**: `.raw/.manifest.json` tracking every ingested source — path, hash, timestamp, which wiki pages it produced. Re-ingest only processes new/changed files.
|
||||
**Current state**: Every `/wiki-ingest` call re-processes everything.
|
||||
**How to add**:
|
||||
- On ingest: compute MD5 hash of source → check manifest → skip if unchanged
|
||||
- On ingest: record `{path, hash, ingested_at, pages_created}` in manifest
|
||||
- On update: re-process if hash changed, merge changes into existing pages
|
||||
|
||||
### 5. Multi-Depth Query Modes
|
||||
**Source**: rvk7895/llm-knowledge-bases
|
||||
**What it is**: 3 query tiers in `/wiki-query`:
|
||||
- **Quick** — hot.md + index.md only (~3 pages read)
|
||||
- **Standard** — full wiki cross-reference + optional web search supplement
|
||||
- **Deep** — parallel sub-agents, each researching a different angle
|
||||
**Current state**: One depth level.
|
||||
**How to add**: `/wiki-query quick <question>`, `/wiki-query deep <question>` flags in SKILL.md.
|
||||
|
||||
### 6. /wiki-ingest Vision Support
|
||||
**Source**: Ar9av/obsidian-wiki
|
||||
**What it is**: Ingest images, screenshots, whiteboard photos by passing the image to a vision-capable model.
|
||||
**How to add**: Detect image extension → read as base64 → pass to Claude with vision prompt asking for transcription/description → treat result as text source → standard ingest pipeline.
|
||||
**Useful for**: Whiteboard photos from meetings, screenshots of web content, diagrams.
|
||||
|
||||
---
|
||||
|
||||
## Tier 3 — Bigger Features Worth Planning
|
||||
|
||||
### 7. /adopt — Import Existing Vault
|
||||
**Source**: heyitsnoah/claudesidian, ballred/obsidian-claude-pkm
|
||||
**What it is**: `/adopt` analyzes an existing Obsidian vault, detects its organization method (PARA, Zettelkasten, LYT, plain), and wraps the LLM Wiki pattern around it without destroying existing structure.
|
||||
**Why it matters**: Currently, users must start fresh. This unlocks adoption by people with existing vaults.
|
||||
**Implementation**: Scan folder structure → classify patterns → generate CLAUDE.md mapping existing folders to wiki roles → non-destructive.
|
||||
|
||||
### 8. Productivity Wrapper (Daily/Weekly Reviews)
|
||||
**Source**: ballred/obsidian-claude-pkm
|
||||
**What it is**: Optional `/daily` and `/weekly` skills that connect goal tracking to the knowledge base.
|
||||
**Could be a separate plugin** rather than bundled into claude-obsidian.
|
||||
**Goal cascade**: 3-Year Vision → Yearly Goals → Projects → Weekly → Daily.
|
||||
|
||||
### 9. Multi-Agent Compatibility (Cursor, Windsurf, Codex)
|
||||
**Source**: Ar9av/obsidian-wiki, kepano/obsidian-skills
|
||||
**What it is**: A `setup.sh` or `/wiki-convert` command that generates `.cursor/rules/`, `AGENTS.md`, `GEMINI.md` equivalents so the wiki skills work in other coding agents.
|
||||
**Note**: kepano already published skills in Agent Skills format — claude-obsidian is already in that format. Just needs the adapter files.
|
||||
|
||||
### 10. Marp Presentation Output
|
||||
**Source**: rvk7895/llm-knowledge-bases, ekadetov/llm-wiki
|
||||
**What it is**: `/wiki-query --slides <topic>` generates a Marp presentation from wiki content, saved to `output/`.
|
||||
**Requires**: `marp-cli` npm package.
|
||||
|
||||
---
|
||||
|
||||
## Tier 4 — Research / Ecosystem Plays
|
||||
|
||||
### 11. obsidian-memory-mcp Integration
|
||||
**Source**: YuNaga224/obsidian-memory-mcp
|
||||
**What it is**: Connect the MCP server that stores Claude's memories as Markdown entities with `[[wikilinks]]` → they appear in Obsidian graph view automatically.
|
||||
**How to add**: Point MEMORY_DIR to the wiki/entities/ directory — entity memory pages become proper wiki pages.
|
||||
|
||||
### 12. obsidian-bases Skill (from kepano)
|
||||
**Source**: kepano/obsidian-skills
|
||||
**What it is**: Teach Claude how to create and edit Obsidian Bases (.base files) for dynamic tables, views, and filters.
|
||||
**Why**: Obsidian Bases is a new core feature — no other LLM Wiki project teaches Claude about it yet.
|
||||
|
||||
### 13. Schema-Emergent Vault Mode
|
||||
**Source**: Ar9av/obsidian-wiki
|
||||
**What it is**: Alternative /wiki mode where the vault structure is not scaffolded upfront but emerges from ingested content. Good for exploratory knowledge building vs. structured domains.
|
||||
**How**: Skip the scaffold step; let wiki-ingest create folders/categories organically based on source content.
|
||||
|
||||
---
|
||||
|
||||
## Competitive Positioning
|
||||
|
||||
After this research, claude-obsidian's unique advantages remain:
|
||||
- **Hot cache** — no one else has this session context mechanism
|
||||
- **Canvas visual layer** — unique in the LLM Wiki category
|
||||
- **/save conversation** — filing chat → wiki is a distinct workflow
|
||||
- **Marketplace polish** — best install experience in category
|
||||
- **Community distribution** (avalonreset-pro)
|
||||
|
||||
The ecosystem is maturing fast. Tier 1 items (URL ingest, auto-commit, defuddle) should ship in v1.3.0 to stay ahead.
|
||||
|
||||
---
|
||||
|
||||
## Implementation Priority
|
||||
|
||||
```
|
||||
v1.3.0 (quick wins):
|
||||
- URL ingestion (#1)
|
||||
- Auto-commit hook (#2)
|
||||
- defuddle integration (#3)
|
||||
|
||||
v1.4.0 (quality):
|
||||
- Delta tracking (#4)
|
||||
- Multi-depth query (#5)
|
||||
|
||||
v1.5.0 (expansion):
|
||||
- Vision ingest (#6)
|
||||
- /adopt command (#7)
|
||||
- Multi-agent compat (#9)
|
||||
|
||||
Future:
|
||||
- Productivity wrapper (#8)
|
||||
- Marp output (#10)
|
||||
- Memory MCP integration (#11)
|
||||
```
|
||||
@@ -0,0 +1,47 @@
|
||||
---
|
||||
type: entity
|
||||
title: "Andrej Karpathy"
|
||||
entity_type: person
|
||||
role: "AI researcher, educator, founder"
|
||||
first_mentioned: "[[LLM Wiki Pattern]]"
|
||||
created: 2026-04-07
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- entity
|
||||
- person
|
||||
- ai-researcher
|
||||
status: mature
|
||||
related:
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Compounding Knowledge]]"
|
||||
- "[[entities/_index]]"
|
||||
sources:
|
||||
---
|
||||
|
||||
# Andrej Karpathy
|
||||
|
||||
AI researcher and educator. Former Director of AI at Tesla, founding member of OpenAI. Known for deep learning education (Neural Networks: Zero to Hero series, micrograd, nanoGPT) and for publishing practical AI patterns.
|
||||
|
||||
---
|
||||
|
||||
## Key Contributions to This Wiki
|
||||
|
||||
Karpathy originated the [[LLM Wiki Pattern]] — the idea of using LLMs to build and maintain a persistent, compounding knowledge base rather than re-deriving knowledge from raw documents on every query.
|
||||
|
||||
He published the idea as an "idea file" — intentionally abstract, designed to be copy-pasted into a Claude Code or similar session and built out collaboratively. The claude-obsidian plugin is a production implementation of this pattern.
|
||||
|
||||
His framing: "The wiki is a persistent, compounding artifact. The cross-references are already there. The contradictions have already been flagged. The synthesis already reflects everything you've read."
|
||||
|
||||
---
|
||||
|
||||
## Notable Quote
|
||||
|
||||
"I thought that I had to reach for fancy RAG, but the LLM has been pretty good at auto-maintaining index files and brief summaries of all documents and reads all the important related data fairly easily at this small scale."
|
||||
|
||||
---
|
||||
|
||||
## Connections
|
||||
|
||||
- [[LLM Wiki Pattern]] — his core contribution to knowledge management
|
||||
- [[Compounding Knowledge]] — the central insight he articulated
|
||||
- claude-obsidian plugin — this repo is a production implementation of his pattern
|
||||
@@ -0,0 +1,61 @@
|
||||
---
|
||||
type: entity
|
||||
title: "Ar9av/obsidian-wiki"
|
||||
created: 2026-04-08
|
||||
updated: 2026-04-08
|
||||
tags:
|
||||
- github-repo
|
||||
- llm-wiki-pattern
|
||||
- multi-agent
|
||||
status: current
|
||||
related:
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[cherry-picks]]"
|
||||
- "[[claude-obsidian-ecosystem]]"
|
||||
sources:
|
||||
- "[[claude-obsidian-ecosystem-research]]"
|
||||
---
|
||||
|
||||
# Ar9av/obsidian-wiki
|
||||
|
||||
**Type**: Claude Code plugin (skill-based)
|
||||
**URL**: https://github.com/Ar9av/obsidian-wiki
|
||||
**Pattern**: Karpathy LLM Wiki
|
||||
**Unique claim**: Works with any AI coding agent via `setup.sh`
|
||||
|
||||
## What It Does
|
||||
|
||||
Framework for AI agents to build and maintain an Obsidian wiki using the Karpathy LLM Wiki pattern. The key differentiator: a single `setup.sh` deploys the skills to 7 different agents simultaneously.
|
||||
|
||||
## Agent Compatibility Matrix
|
||||
|
||||
| Agent | Bootstrap | Skills Dir |
|
||||
|-------|-----------|-----------|
|
||||
| Claude Code | CLAUDE.md | `.claude/skills/` |
|
||||
| Cursor | `.cursor/rules/obsidian-wiki.mdc` | `.cursor/skills/` |
|
||||
| Windsurf | `.windsurf/rules/` | `.windsurf/skills/` |
|
||||
| Codex (OpenAI) | AGENTS.md | `~/.codex/skills/` |
|
||||
| Gemini/Antigravity | GEMINI.md | `~/.gemini/antigravity/skills/` |
|
||||
| OpenClaw | AGENTS.md | `.agents/skills/` |
|
||||
| GitHub Copilot | `.github/copilot-instructions.md` | — |
|
||||
|
||||
## Key Innovations
|
||||
|
||||
### Delta Tracking Manifest
|
||||
`.manifest.json` tracks every ingested source: path, hash, timestamp, which wiki pages produced. Only processes new/changed files. Solves the "re-ingest everything" problem.
|
||||
|
||||
### 4-Stage Pipeline
|
||||
1. **Ingest** — reads source (PDF, JSONL, text, conversation exports, images)
|
||||
2. **Extract** — pulls concepts, entities, claims, relationships, open questions
|
||||
3. **Resolve** — merges new knowledge against existing wiki (no duplication)
|
||||
4. **Schema** — structure emerges from sources, not predefined
|
||||
|
||||
### Vision Support
|
||||
Images, screenshots, whiteboard photos ingestable with vision-capable model. Each page gets 1-2 sentence `summary:` in frontmatter for preview without opening.
|
||||
|
||||
## Cherry-Picks for claude-obsidian
|
||||
|
||||
- [[cherry-picks#4. Delta Tracking Manifest]]
|
||||
- [[cherry-picks#6. /wiki-ingest Vision Support]]
|
||||
- [[cherry-picks#9. Multi-Agent Compatibility (Cursor, Windsurf, Codex)]]
|
||||
- [[cherry-picks#13. Schema-Emergent Vault Mode]]
|
||||
@@ -0,0 +1,74 @@
|
||||
---
|
||||
type: entity
|
||||
title: "Claude SEO"
|
||||
created: 2026-04-14
|
||||
updated: 2026-04-15
|
||||
tags:
|
||||
- entity
|
||||
- project
|
||||
- claude-code
|
||||
- seo
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[Pro Hub Challenge]]"
|
||||
- "[[2026-04-14-claude-seo-v190-session]]"
|
||||
- "[[Semantic Topic Clustering]]"
|
||||
- "[[Search Experience Optimization]]"
|
||||
- "[[SEO Drift Monitoring]]"
|
||||
- "[[E-commerce SEO]]"
|
||||
- "[[2026-04-15-slides-and-release-session]]"
|
||||
- "[[2026-04-15-release-report-session]]"
|
||||
---
|
||||
|
||||
# Claude SEO
|
||||
|
||||
A Tier 4 Claude Code skill for comprehensive SEO analysis across all industries. Repository: [AgriciDaniel/claude-seo](https://github.com/AgriciDaniel/claude-seo)
|
||||
|
||||
## Current State (v1.9.0 — released April 15, 2026)
|
||||
|
||||
- **23 skills** (20 core + 3 extensions: DataForSEO, Firecrawl, Banana)
|
||||
- **17 subagents** (15 core + 2 extension agents)
|
||||
- **30 Python scripts** (28 tracked + 2 dev-only)
|
||||
- **Architecture**: 3-layer (directive, orchestration, execution)
|
||||
- **Entry point**: `/seo [command] [url]`
|
||||
- **GitHub release**: [v1.9.0](https://github.com/AgriciDaniel/claude-seo/releases/tag/v1.9.0) — PDF report attached
|
||||
- **Slides**: `claude-seo-slides/v190.html` — 15-slide community presentation deck
|
||||
- **Contributors**: 6 submissions, 5 integrated (Lutfiya Miller, Florian Schmitz, Dan Colta, Matej Marjanovic, Chris Muller)
|
||||
|
||||
## Key Commands
|
||||
|
||||
| Category | Commands |
|
||||
|----------|----------|
|
||||
| Analysis | audit, page, technical, content, schema, images, geo |
|
||||
| Planning | plan, cluster, sxo, programmatic, competitor-pages |
|
||||
| Monitoring | drift baseline, drift compare, drift history |
|
||||
| Local | local, maps |
|
||||
| International | hreflang (with cultural profiles) |
|
||||
| E-commerce | ecommerce |
|
||||
| Data | google, backlinks, dataforseo |
|
||||
| Generation | sitemap, image-gen |
|
||||
|
||||
## Version History
|
||||
|
||||
| Version | Date | Key Addition |
|
||||
|---------|------|-------------|
|
||||
| v1.9.0 | 2026-04-15 | Pro Hub Challenge: cluster, SXO, drift, ecommerce, cost guardrails, cultural profiles. GitHub release + PDF report + 15-slide deck. |
|
||||
| v1.8.2 | 2026-04-10 | Ukrainian localization, CI fixes, version sync |
|
||||
| v1.8.1 | 2026-04-06 | Google Images SERP, image optimization |
|
||||
| v1.8.0 | 2026-03-31 | Free backlink data (Moz, Bing, Common Crawl) |
|
||||
| v1.7.0 | 2026-03-28 | Google SEO APIs (GSC, PageSpeed, CrUX, GA4) |
|
||||
|
||||
## Ecosystem
|
||||
|
||||
- [[Claude SEO]] — SEO analysis (this project)
|
||||
- Claude Blog — companion blog engine, consumes SEO findings
|
||||
- Claude Banana — AI image generation, bundled as extension
|
||||
- AI Marketing Claude — community marketing suite by Zubair Trabzada
|
||||
|
||||
## Security Posture (v1.9.0 audit)
|
||||
|
||||
- **Score**: 85/100 (Grade B+)
|
||||
- **SSRF protection**: validate_url() + fetch_page.py DNS resolution
|
||||
- **SQL**: all queries parameterized
|
||||
- **Cost guardrails**: threshold approval, daily limits, file locking, audit trail
|
||||
- **Pre-existing debt**: validate_url DNS rebinding gap, install script injection, OAuth file permissions
|
||||
@@ -0,0 +1,59 @@
|
||||
---
|
||||
type: entity
|
||||
title: "Claudian (YishenTu/claudian)"
|
||||
created: 2026-04-08
|
||||
updated: 2026-04-08
|
||||
tags:
|
||||
- github-repo
|
||||
- native-obsidian-plugin
|
||||
- embedded-ai
|
||||
status: current
|
||||
related:
|
||||
- "[[cherry-picks]]"
|
||||
- "[[claude-obsidian-ecosystem]]"
|
||||
sources:
|
||||
- "[[claude-obsidian-ecosystem-research]]"
|
||||
---
|
||||
|
||||
# Claudian (YishenTu/claudian)
|
||||
|
||||
**Type**: Native Obsidian plugin (TypeScript, embedded Claude Code/Codex)
|
||||
**URL**: https://github.com/YishenTu/claudian
|
||||
**Install**: BRAT or manual (not yet in community store)
|
||||
|
||||
## What It Does
|
||||
|
||||
Embeds Claude Code (or Codex CLI) directly inside Obsidian as a sidebar chat. The vault becomes the agent's working directory — all Claude Code tools work natively inside Obsidian.
|
||||
|
||||
## Key Features
|
||||
|
||||
### Inline Edit with Word-Level Diff
|
||||
Select text in a note + hotkey → Claude proposes edit with word-level diff preview → one-click apply. Best-in-class inline editing in the Obsidian AI ecosystem.
|
||||
|
||||
### Plan Mode (Shift+Tab)
|
||||
Agent explores and designs before implementing. Presents a plan for approval before any file changes. Mirrors Claude Code's own plan mode.
|
||||
|
||||
### @mention System
|
||||
Type `@` to reference:
|
||||
- Vault files
|
||||
- Sub-agents
|
||||
- MCP servers
|
||||
- Files in external directories (outside vault)
|
||||
|
||||
### Instruction Mode (#)
|
||||
Add refined custom instructions directly from chat input — persisted for the session.
|
||||
|
||||
### MCP Server Integration
|
||||
Connect external tools via stdio, SSE, or HTTP. Claude manages vault MCP in-app; Codex uses CLI-managed config.
|
||||
|
||||
### Multi-Tab Conversations
|
||||
Multiple chat tabs, conversation history, fork, resume, compact mode.
|
||||
|
||||
## Privacy
|
||||
- No telemetry
|
||||
- Settings stored in `vault/.claudian/`
|
||||
- Claude files in `vault/.claude/`
|
||||
- Transcripts in `~/.claude/projects/`
|
||||
|
||||
## Relevance to claude-obsidian
|
||||
Claudian is a native plugin (different category) but its Plan Mode, @mention, and inline edit patterns could inspire new features in claude-obsidian skills — particularly for the canvas and wiki-query workflows.
|
||||
@@ -0,0 +1,53 @@
|
||||
---
|
||||
type: entity
|
||||
title: "Nexus (ProfSynapse/claudesidian-mcp)"
|
||||
created: 2026-04-08
|
||||
updated: 2026-04-08
|
||||
tags:
|
||||
- github-repo
|
||||
- obsidian-plugin
|
||||
- mcp-server
|
||||
- native-plugin
|
||||
status: current
|
||||
related:
|
||||
- "[[cherry-picks]]"
|
||||
- "[[claude-obsidian-ecosystem]]"
|
||||
sources:
|
||||
- "[[claude-obsidian-ecosystem-research]]"
|
||||
---
|
||||
|
||||
# Nexus (formerly Claudesidian MCP)
|
||||
|
||||
**Type**: Native Obsidian plugin + MCP bridge
|
||||
**URL**: https://github.com/ProfSynapse/claudesidian-mcp
|
||||
**Current name**: Nexus MCP for Obsidian
|
||||
**Install**: `.obsidian/plugins/nexus/`
|
||||
|
||||
## What It Does
|
||||
|
||||
Full Obsidian plugin with two modes:
|
||||
1. **Native chat inside Obsidian** — connect any AI provider
|
||||
2. **MCP bridge** — expose vault to Claude Desktop, Claude Code, Codex CLI, Gemini CLI, Cursor, Cline
|
||||
|
||||
## Key Features
|
||||
|
||||
- **Workspace memory** — persistent context across sessions stored as JSONL, automatically included in Obsidian Sync
|
||||
- **Task management** — projects, tasks, blockers, dependencies tracked within vault
|
||||
- **Semantic search** — search notes + past conversations by meaning
|
||||
- **Inline editing** — edit selected text in notes
|
||||
- **PDF + audio → Markdown** — conversion on right-click or auto-on-add
|
||||
- **Web page capture** — open URL in Obsidian, save as Markdown/PNG/PDF
|
||||
- **Mobile support** — native chat works on iOS/Android
|
||||
- **Two-tool architecture** — dedicated tools for read vs write actions
|
||||
|
||||
## Storage Architecture
|
||||
|
||||
Data stored as JSONL files in `.obsidian/plugins/nexus/data/`. This is included in Obsidian Sync automatically (unlike the `.nexus/` folder of v1). SQLite cache is local-only, rebuilt from JSONL on each device.
|
||||
|
||||
## Relevance to claude-obsidian
|
||||
|
||||
Nexus is in a different category — it's a native TypeScript Obsidian plugin, not a Claude Code skill plugin. The two don't compete directly, but its workspace memory and task management patterns are cherry-pickable.
|
||||
|
||||
## Cherry-Picks for claude-obsidian
|
||||
|
||||
- [[cherry-picks#11. obsidian-memory-mcp Integration]] (different implementation, same concept)
|
||||
@@ -0,0 +1,43 @@
|
||||
---
|
||||
type: meta
|
||||
title: "Entities Index"
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- meta
|
||||
- index
|
||||
- entity
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[index]]"
|
||||
- "[[Andrej Karpathy]]"
|
||||
- "[[hot]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
---
|
||||
|
||||
# Entities Index
|
||||
|
||||
Navigation: [[index]] | [[concepts/_index|Concepts]] | [[sources/_index|Sources]]
|
||||
|
||||
All entity pages — people, organizations, products, and tools.
|
||||
|
||||
---
|
||||
|
||||
## People
|
||||
|
||||
- [[Andrej Karpathy]] — AI researcher, educator; originated the LLM Wiki pattern
|
||||
|
||||
---
|
||||
|
||||
## Organizations
|
||||
|
||||
<!-- Add organization pages here -->
|
||||
|
||||
---
|
||||
|
||||
## Products & Tools
|
||||
|
||||
<!-- Add tool and product pages here -->
|
||||
|
||||
---
|
||||
|
||||
## Add new entities here as they are identified during ingests.
|
||||
@@ -0,0 +1,66 @@
|
||||
---
|
||||
type: entity
|
||||
title: "ballred/obsidian-claude-pkm"
|
||||
created: 2026-04-08
|
||||
updated: 2026-04-08
|
||||
tags:
|
||||
- github-repo
|
||||
- llm-wiki-pattern
|
||||
- pkm
|
||||
- productivity
|
||||
status: current
|
||||
related:
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[cherry-picks]]"
|
||||
- "[[claude-obsidian-ecosystem]]"
|
||||
sources:
|
||||
- "[[claude-obsidian-ecosystem-research]]"
|
||||
---
|
||||
|
||||
# ballred/obsidian-claude-pkm
|
||||
|
||||
**Type**: Claude Code plugin (skill-based PKM system)
|
||||
**URL**: https://github.com/ballred/obsidian-claude-pkm
|
||||
**Version**: 3.1
|
||||
**Tagline**: "Not another PKM starter kit. An execution system."
|
||||
|
||||
## What It Does
|
||||
|
||||
Connects a 3-year vision cascade to daily task execution, using Claude as an accountability partner. Every layer is linked — daily notes surface the weekly ONE Big Thing, which links to active projects, which link to yearly goals.
|
||||
|
||||
## Goal Cascade
|
||||
|
||||
```
|
||||
3-Year Vision → Yearly Goals → Projects → Monthly Goals → Weekly Review → Daily Tasks
|
||||
```
|
||||
|
||||
Each layer has a dedicated skill: `/goal-tracking`, `/project`, `/monthly`, `/weekly`, `/daily`, `/review`.
|
||||
|
||||
## Key Innovations
|
||||
|
||||
### Auto-Commit via PostToolUse Hook
|
||||
Every Write/Edit tool call triggers `git add -A && git commit` automatically. Vault is always versioned.
|
||||
|
||||
### /adopt Command
|
||||
Scans an existing Obsidian vault, detects its organization method (PARA, Zettelkasten, LYT, plain folders), maps folders interactively to the PKM layers, generates config files. Non-destructive.
|
||||
|
||||
### 4 Specialized Agents with Memory
|
||||
- `goal-aligner` — audits activity vs. stated goals, flags misalignment
|
||||
- `weekly-reviewer` — facilitates 3-phase weekly review, learns reflection style
|
||||
- `note-organizer` — fixes broken links, consolidates duplicates
|
||||
- `inbox-processor` — GTD-style inbox processing
|
||||
|
||||
Uses `memory: project` so agents remember patterns across sessions.
|
||||
|
||||
### Productivity Coach Output Style
|
||||
`/output-style coach` transforms Claude into an accountability partner — challenges assumptions, asks powerful questions, points out goal-action misalignment.
|
||||
|
||||
## Architecture
|
||||
|
||||
Zero dependencies (bash + markdown only). Path-specific rules loaded contextually. Session init surfaces ONE Big Thing, active project count, days since last review.
|
||||
|
||||
## Cherry-Picks for claude-obsidian
|
||||
|
||||
- [[cherry-picks#2. Auto-Commit PostToolUse Hook]]
|
||||
- [[cherry-picks#7. /adopt — Import Existing Vault]]
|
||||
- [[cherry-picks#8. Productivity Wrapper (Daily/Weekly Reviews)]]
|
||||
@@ -0,0 +1,62 @@
|
||||
---
|
||||
type: entity
|
||||
title: "kepano/obsidian-skills"
|
||||
created: 2026-04-08
|
||||
updated: 2026-04-08
|
||||
tags:
|
||||
- github-repo
|
||||
- official
|
||||
- agent-skills
|
||||
- obsidian-creator
|
||||
status: current
|
||||
related:
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[cherry-picks]]"
|
||||
- "[[claude-obsidian-ecosystem]]"
|
||||
sources:
|
||||
- "[[claude-obsidian-ecosystem-research]]"
|
||||
---
|
||||
|
||||
# kepano/obsidian-skills
|
||||
|
||||
**Type**: Agent Skills (Agent Skills standard)
|
||||
**URL**: https://github.com/kepano/obsidian-skills
|
||||
**Author**: **Linus Kepano** — creator of Obsidian + Minimal theme
|
||||
**Install**: `/plugin marketplace add kepano/obsidian-skills`
|
||||
|
||||
## Why This Matters
|
||||
|
||||
This repo is from the creator of Obsidian. It:
|
||||
1. Validates that the Agent Skills standard is the right format for Obsidian AI tools
|
||||
2. Provides the canonical reference for how to teach Claude about Obsidian-specific syntax
|
||||
3. Covers Obsidian Bases — a new core Obsidian feature that no other AI project supports yet
|
||||
|
||||
## Skills
|
||||
|
||||
| Skill | What It Teaches Claude |
|
||||
|-------|----------------------|
|
||||
| `obsidian-markdown` | Full Obsidian Flavored Markdown: wikilinks, embeds, callouts, properties, tags |
|
||||
| `obsidian-bases` | Obsidian Bases (.base files): views, filters, formulas, summaries |
|
||||
| `json-canvas` | JSON Canvas spec: nodes, edges, groups, connections |
|
||||
| `obsidian-cli` | Vault management, plugin/theme dev via Obsidian CLI |
|
||||
| `defuddle` | Extract clean Markdown from web pages — removes ads, nav, clutter |
|
||||
|
||||
## defuddle
|
||||
|
||||
The `defuddle` skill wraps `defuddle-cli`. When ingesting web content, running defuddle first:
|
||||
- Strips ads, navigation, footers
|
||||
- Reduces token usage ~40-60% on typical web pages
|
||||
- Produces cleaner Markdown that fits better in context window
|
||||
|
||||
This is a direct cherry-pick for claude-obsidian's ingest pipeline.
|
||||
|
||||
## Multi-Platform
|
||||
|
||||
Works with Claude Code, Codex CLI, and OpenCode out of the box.
|
||||
|
||||
## Cherry-Picks for claude-obsidian
|
||||
|
||||
- [[cherry-picks#1. URL Ingestion in /wiki-ingest]] (pair with defuddle)
|
||||
- [[cherry-picks#3. defuddle Web Cleaning Skill]]
|
||||
- [[cherry-picks#12. obsidian-bases Skill (from kepano)]]
|
||||
- [[cherry-picks#9. Multi-Agent Compatibility]] (format already compatible)
|
||||
@@ -0,0 +1,61 @@
|
||||
---
|
||||
type: entity
|
||||
title: "rvk7895/llm-knowledge-bases"
|
||||
created: 2026-04-08
|
||||
updated: 2026-04-08
|
||||
tags:
|
||||
- github-repo
|
||||
- llm-wiki-pattern
|
||||
- deep-research
|
||||
status: current
|
||||
related:
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[cherry-picks]]"
|
||||
- "[[claude-obsidian-ecosystem]]"
|
||||
- "[[Andrej Karpathy]]"
|
||||
sources:
|
||||
- "[[claude-obsidian-ecosystem-research]]"
|
||||
---
|
||||
|
||||
# rvk7895/llm-knowledge-bases
|
||||
|
||||
**Type**: Claude Code plugin (Marketplace)
|
||||
**URL**: https://github.com/rvk7895/llm-knowledge-bases
|
||||
**Install**: `/plugin marketplace add rvk7895/llm-knowledge-bases`
|
||||
|
||||
## What It Does
|
||||
|
||||
Turns raw research material into an LLM-maintained Obsidian wiki with multi-depth querying and rich output formats. Adds a deep research pipeline with parallel agents on top of the Karpathy pattern.
|
||||
|
||||
## Key Innovations
|
||||
|
||||
### 3-Depth Query System
|
||||
- **Quick** — answers from wiki indexes and summaries only (minimal reads)
|
||||
- **Standard** — cross-references full wiki, supplements with web search
|
||||
- **Deep** — multi-agent parallel web search pipeline
|
||||
|
||||
### Output Formats
|
||||
Beyond Markdown: Marp slides, matplotlib charts. All outputs saved to `output/` and optionally filed back into wiki.
|
||||
|
||||
### Skills Set
|
||||
| Skill | Purpose |
|
||||
|-------|---------|
|
||||
| `/kb-init` | One-time setup |
|
||||
| `/kb compile` | Raw → wiki |
|
||||
| `/kb query` | Query with depth |
|
||||
| `/kb lint` | Health check |
|
||||
| `/kb evolve` | Maintenance pass |
|
||||
| `/research <topic>` | Structured research outline |
|
||||
| `/research-deep` | Parallel agents per outline item |
|
||||
| `/research-report` | Compile deep results → Markdown |
|
||||
|
||||
### X/Twitter Integration
|
||||
Via Smaug tool (`npm install -g @steipete/bird`). Ingests tweets, threads, bookmarks from X/Twitter by pasting URL. Uses session cookies (read-only, personal use).
|
||||
|
||||
## Attribution
|
||||
Built on Karpathy pattern + Weizhena's Deep Research skills adapted for the research pipeline.
|
||||
|
||||
## Cherry-Picks for claude-obsidian
|
||||
|
||||
- [[cherry-picks#5. Multi-Depth Query Modes]]
|
||||
- [[cherry-picks#10. Marp Presentation Output]]
|
||||
@@ -0,0 +1,115 @@
|
||||
---
|
||||
type: fold
|
||||
title: "Fold k3 - 2026-04-23 to 2026-04-24 - n8"
|
||||
fold_id: "fold-k3-from-2026-04-23-to-2026-04-24-n8"
|
||||
batch_exponent: 3
|
||||
n: 8
|
||||
entry_count: 8
|
||||
earliest: "2026-04-23"
|
||||
latest: "2026-04-24"
|
||||
entry_range:
|
||||
from: "2026-04-23"
|
||||
to: "2026-04-24"
|
||||
created: "2026-04-24"
|
||||
updated: "2026-04-24"
|
||||
tags:
|
||||
- meta
|
||||
- fold
|
||||
- "fold/k3"
|
||||
status: mature
|
||||
children:
|
||||
- date: "2026-04-24"
|
||||
op: "save"
|
||||
title: "v1.6.0 closeout (Teams, chair-led)"
|
||||
page: "[[2026-04-24-v1.6.0-release-session]]"
|
||||
page_missing: false
|
||||
- date: "2026-04-24"
|
||||
op: "save"
|
||||
title: "DragonScale Phase 4 - boundary-first autoresearch shipped (v1.6.0)"
|
||||
page: "[[DragonScale Memory]]"
|
||||
page_missing: false
|
||||
- date: "2026-04-24"
|
||||
op: "save"
|
||||
title: "DragonScale Phase 3.5 - cross-phase hardening to v1.5.0"
|
||||
page: "[[DragonScale Memory]]"
|
||||
page_missing: false
|
||||
- date: "2026-04-24"
|
||||
op: "save"
|
||||
title: "DragonScale Phase 3 - semantic tiling MVP"
|
||||
page: "[[DragonScale Memory]]"
|
||||
page_missing: false
|
||||
- date: "2026-04-23"
|
||||
op: "save"
|
||||
title: "DragonScale Phase 2 - deterministic page addresses MVP"
|
||||
page: "[[DragonScale Memory]]"
|
||||
page_missing: false
|
||||
- date: "2026-04-23"
|
||||
op: "save"
|
||||
title: "DragonScale Phase 1 - wiki-fold skill shipped"
|
||||
page: "[[wiki-fold]]"
|
||||
page_missing: false
|
||||
- date: "2026-04-23"
|
||||
op: "save"
|
||||
title: "DragonScale Memory v0.2 - post-adversarial-review"
|
||||
page: "[[DragonScale Memory]]"
|
||||
page_missing: false
|
||||
- date: "2026-04-23"
|
||||
op: "save"
|
||||
title: "DragonScale Memory - Phase 0 design doc (proposed)"
|
||||
page: "[[DragonScale Memory]]"
|
||||
page_missing: false
|
||||
related:
|
||||
- "[[DragonScale Memory]]"
|
||||
- "[[log]]"
|
||||
- "[[index]]"
|
||||
---
|
||||
|
||||
Level-3 fold of 8 log entries spanning 2026-04-23 to 2026-04-24. Dominant themes: staged DragonScale implementation, adversarial review gates, release and test hardening.
|
||||
|
||||
## Child Entries
|
||||
|
||||
| Date | Op | Title | Page | Summary (extractive) |
|
||||
|---|---|---|---|---|
|
||||
| 2026-04-24 | save | v1.6.0 closeout (Teams, chair-led) | [[2026-04-24-v1.6.0-release-session]] | Release closeout recorded new release-session, boundary-frontier, and DragonScale guide artifacts; all 11 verifier items were applied and `make test` passed. |
|
||||
| 2026-04-24 | save | DragonScale Phase 4 - boundary-first autoresearch shipped (v1.6.0) | [[DragonScale Memory]] | Boundary-first autoresearch shipped as opt-in topic selection, surfacing top-5 frontier pages when `/autoresearch` has no topic. |
|
||||
| 2026-04-24 | save | DragonScale Phase 3.5 - cross-phase hardening to v1.5.0 | [[DragonScale Memory]] | Cross-phase audit hardening resolved 10 hold-ship items, added a reproducible test harness, and version-bumped plugin metadata to 1.5.0. |
|
||||
| 2026-04-24 | save | DragonScale Phase 3 - semantic tiling MVP | [[DragonScale Memory]] | Semantic tiling shipped as opt-in embedding-based duplicate detection with `scripts/tiling-check.py`, threshold state, lint wiring, and a final 10/10 accept verdict. |
|
||||
| 2026-04-23 | save | DragonScale Phase 2 - deterministic page addresses MVP | [[DragonScale Memory]] | Deterministic page addresses shipped with `c-NNNNNN` allocation, flock-guarded concurrency, lint checks, manifest idempotency, and an 8/8 accept verdict. |
|
||||
| 2026-04-23 | save | DragonScale Phase 1 - wiki-fold skill shipped | [[wiki-fold]] | Wiki-fold shipped as a flat extractive rollup skill with dry-run stdout default, deterministic fold IDs, duplicate detection, and a 7/7 accept verdict. |
|
||||
| 2026-04-23 | save | DragonScale Memory v0.2 - post-adversarial-review | [[DragonScale Memory]] | The v0.2 concept revision weakened claims, added operational policies and provenance tags, and reached 7/7 accepted after one surgical fix. |
|
||||
| 2026-04-23 | save | DragonScale Memory - Phase 0 design doc (proposed) | [[DragonScale Memory]] | The Phase 0 design proposed four memory-layer mechanisms and recorded verified primary sources for dragon curve, paperfolding, LSM trees, MemGPT, and Anthropic prompt caching. |
|
||||
|
||||
## Key Outcomes
|
||||
|
||||
- The Phase 0 design proposed four memory-layer mechanisms: fold operator, content-addressable page paths, semantic tiling lint, and boundary-first autoresearch scoring (from 2026-04-23 Phase 0 design doc entry).
|
||||
- The v0.2 concept revision changed load-bearing claims after adversarial review and reached 7/7 accepted after one surgical fix on tagging-scope language (from 2026-04-23 post-adversarial-review entry).
|
||||
- The wiki-fold skill shipped with dry-run stdout, deterministic `fold_id`, duplicate-range detection, and a final 7/7 accept verdict (from 2026-04-23 Phase 1 entry).
|
||||
- Deterministic page addresses shipped with `c-NNNNNN` format, flock-guarded allocation, counter recovery, manifest-preserved path-to-address mapping, and an 8/8 accept verdict (from 2026-04-23 Phase 2 entry).
|
||||
- Semantic tiling shipped with opt-in ollama `nomic-embed-text` duplicate detection, conservative default bands, cache invalidation, locking, scale guards, and final 10/10 accept (from 2026-04-24 Phase 3 entry).
|
||||
- Cross-phase hardening to v1.5.0 resolved 10 hold-ship items and added tests covering 12 shell assertions plus 18 python assertions (from 2026-04-24 Phase 3.5 entry).
|
||||
- Boundary-first autoresearch shipped as v1.6.0 and the log records that all four DragonScale mechanisms are now shipped and opt-in (from 2026-04-24 Phase 4 entry).
|
||||
- The v1.6.0 closeout applied all 11 verifier items across 6 files and `make test` passed with 12 shell assertions, 18 tiling assertions, and 44 boundary-score assertions (from 2026-04-24 closeout entry).
|
||||
|
||||
## Cross-entry Themes
|
||||
|
||||
- DragonScale progressed by gated phases from design proposal to shipped opt-in mechanisms (supported by: 2026-04-23 Phase 0, 2026-04-23 Phase 1, 2026-04-23 Phase 2, 2026-04-24 Phase 3, 2026-04-24 Phase 4 entries).
|
||||
- Adversarial review was used as a repeated acceptance gate, with explicit accept counts recorded in concept revision, wiki-fold, address assignment, semantic tiling, Phase 4, and closeout entries (supported by: 2026-04-23 post-adversarial-review, 2026-04-23 Phase 1, 2026-04-23 Phase 2, 2026-04-24 Phase 3, 2026-04-24 Phase 4, 2026-04-24 closeout entries).
|
||||
- Release hardening added tests, docs, changelog, version metadata, and guide updates around the DragonScale mechanisms (supported by: 2026-04-24 Phase 3.5, 2026-04-24 Phase 4, 2026-04-24 closeout entries).
|
||||
|
||||
## Contradictions or Corrections
|
||||
|
||||
- The 2026-04-24 closeout entry records applied corrections: install-guide wording changed from hierarchical log folds to flat extractive log folds, boundary-frontier provenance was corrected to `--top 7`, and hot.md tag wording was corrected from local tag only to local commits only. Resolution: applied in the closeout entry.
|
||||
|
||||
## Child Pages
|
||||
|
||||
- [[2026-04-24-v1.6.0-release-session]]
|
||||
- [[boundary-frontier-2026-04-24]]
|
||||
- [[DragonScale Memory]]
|
||||
- [[wiki-fold]]
|
||||
- [[fold-template]]
|
||||
|
||||
## Related
|
||||
|
||||
- [[DragonScale Memory]] - fold-operator spec and DragonScale mechanism record
|
||||
- [[log]] - source entries
|
||||
- [[index]] - vault catalog
|
||||
@@ -0,0 +1,100 @@
|
||||
---
|
||||
type: meta
|
||||
title: "Getting Started"
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- meta
|
||||
- onboarding
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[index]]"
|
||||
- "[[overview]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Wiki vs RAG]]"
|
||||
- "[[Wiki Map]]"
|
||||
---
|
||||
|
||||
# Getting Started with claude-obsidian
|
||||
|
||||
Welcome. This vault is your compounding knowledge base — a persistent second brain built with Claude and Obsidian.
|
||||
|
||||
Every source you add gets processed into 8–15 cross-referenced wiki pages. Every question you ask pulls from everything that's been ingested. Knowledge compounds like interest.
|
||||
|
||||
---
|
||||
|
||||
## Three-Step Quick Start
|
||||
|
||||
### 1. Drop a source
|
||||
|
||||
Put any document into the `.raw/` folder:
|
||||
- PDFs, markdown files, transcripts, articles
|
||||
- Or paste a URL and ask Claude to fetch it
|
||||
|
||||
### 2. Ingest it
|
||||
|
||||
Tell Claude in any Claude Code session:
|
||||
|
||||
```
|
||||
ingest [filename]
|
||||
```
|
||||
|
||||
Claude reads the source, creates 8–15 wiki pages under `wiki/`, cross-references everything, and updates `wiki/index.md`, `wiki/log.md`, and `wiki/hot.md`.
|
||||
|
||||
### 3. Ask questions
|
||||
|
||||
```
|
||||
what do you know about [topic]?
|
||||
```
|
||||
|
||||
Claude reads the hot cache, scans the index, drills into relevant pages, and gives you a synthesized answer — citing specific wiki pages, not training data.
|
||||
|
||||
---
|
||||
|
||||
## How the Hot Cache Works
|
||||
|
||||
`wiki/hot.md` is a ~500-word summary of recent vault context. It loads automatically at the start of every session (via the SessionStart hook).
|
||||
|
||||
You don't need to recap. Claude starts every session knowing what you've been working on.
|
||||
|
||||
Update it manually at any time: `update hot cache`
|
||||
|
||||
---
|
||||
|
||||
## Your First Ingest — Walkthrough
|
||||
|
||||
1. Create a file in `.raw/` — copy a transcript, paste an article, or save a PDF
|
||||
2. Open Claude Code in this vault folder
|
||||
3. Type: `ingest [your-filename]`
|
||||
4. Watch the wiki grow — Claude will report which pages it created
|
||||
5. Open `wiki/index.md` — you'll see the new pages listed
|
||||
6. Open Graph View in Obsidian — a new cluster of connected nodes appears
|
||||
|
||||
After 3–5 ingests, the graph starts to look like a real knowledge network. Cross-references emerge automatically.
|
||||
|
||||
---
|
||||
|
||||
## Key Commands
|
||||
|
||||
| You say | Claude does |
|
||||
|---------|-------------|
|
||||
| `ingest [file]` | Creates 8–15 wiki pages from a source |
|
||||
| `what do you know about X?` | Queries the wiki, cites pages |
|
||||
| `/save` | Files this conversation as a wiki note |
|
||||
| `/autoresearch [topic]` | Searches the web, ingests results autonomously |
|
||||
| `lint the wiki` | Health check — finds orphans, gaps, stale links |
|
||||
| `update hot cache` | Refreshes the session context summary |
|
||||
|
||||
---
|
||||
|
||||
## Navigate the Vault
|
||||
|
||||
- **[[Wiki Map]]** — visual Fibonacci graph of all wiki pages
|
||||
- **[[index]]** — master catalog, all pages by type
|
||||
- **[[overview]]** — executive summary of vault contents
|
||||
- **[[LLM Wiki Pattern]]** — the pattern this vault is built on
|
||||
- **[[Wiki vs RAG]]** — why a wiki beats RAG at human scale
|
||||
- **[[dashboard]]** — live Dataview queries (requires Dataview plugin)
|
||||
|
||||
---
|
||||
|
||||
*Built on the [LLM Wiki pattern](https://github.com/karpathy) by Andrej Karpathy.*
|
||||
+92
@@ -0,0 +1,92 @@
|
||||
---
|
||||
type: meta
|
||||
title: "Hot Cache"
|
||||
updated: 2026-05-17T04:30:00
|
||||
tags:
|
||||
- meta
|
||||
- hot-cache
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[index]]"
|
||||
- "[[log]]"
|
||||
- "[[Wiki Map]]"
|
||||
- "[[getting-started]]"
|
||||
- "[[DragonScale Memory]]"
|
||||
---
|
||||
|
||||
# Recent Context
|
||||
|
||||
Navigation: [[index]] | [[log]] | [[overview]]
|
||||
|
||||
## Last Updated
|
||||
|
||||
2026-05-17 (very late, post-polish): **v1.7.1 patch + polish slice shipped locally** (branch `v1.7.0-compound-vault`, still NOT pushed). All 1 BLOCKER + 6 HIGH findings closed; then verifier agent re-pass surfaced 2 MEDIUM + 3 LOW polish items, all closed in `c2d7575`. Final verifier verdict: 0/0/0/0 SHIP. Score: 100/100 on the v1.7.1 patch dimensions (plan fidelity, behavioral correctness, test health, internal consistency, constraint honor, defect introduction, kernel application). 8 commits landed in this resumption session: `ca68bb6` (Fix 1+6 BLOCKER B1 + H6 — contextual-prefix `--allow-egress` flag default-off + `bin/setup-retrieve.sh` consent prompt + `skills/wiki-retrieve/SKILL.md` Data Privacy callout, mirror of `tiling-check.py:351` `--allow-remote-ollama` precedent), `4837d4f` (Fix 2 H1 — setup-retrieve exit 5 + 3-option recovery hint on Stage 1 failure), `7e1f187` (Fix 3 H2 — `make clean-test-state` extended to v1.7 artifacts), `7120970` (Fix 4 H3 — PostToolUse hook captures LOCK_RC directly, not via pipeline; defers commit on script error OR locks held), `722ac97` (Fix 5 H5 — `detect-transport.sh` `json_escape()` helper via `python3 json.dumps`), `3ea443f` (Fix 7 H4 — new `agents/verifier.md` read-only pre-commit specialist + CLAUDE.md reference), and the cross-cutting closeout `822c80a` (version bump 1.7.0 → 1.7.1, CHANGELOG entry, audit doc updated with §10.2 SHAs + v1.7.1 closeout block, audit benchmark scripts promoted to tracked files). `make test` ran 7/7 green after every fix. End-to-end verifications: `python3 scripts/contextual-prefix.py --peek` returns `tier=synthetic` even with `ANTHROPIC_API_KEY` set (default-deny works); `--allow-egress` correctly flips it; `echo "" | bash bin/setup-retrieve.sh` aborts at the consent prompt; `bash scripts/wiki-lock.sh acquire ...` then hook trigger correctly defers auto-commit. **Next step**: ask user whether to push + tag `v1.7.1`. Do NOT push without explicit go.
|
||||
|
||||
2026-05-17 (late): **v1.7.0 full audit complete; v1.7.1 fixes plan ready**. Branch `v1.7.0-compound-vault` still local-only (no push, no tag). The audit was demanded as "THROUGH and FULL on AUDIT following /best-practices" with EVERYTHING scope. Result: **31 findings (1 BLOCKER + 6 HIGH + 14 MEDIUM + 10 LOW)** in `docs/audits/v1.7.0-audit-2026-05-17.md` (481 lines). **BLOCKER**: `scripts/contextual-prefix.py:252-258` data-egress consent gap — `pick_prefix_tier()` silently sends wiki page bodies to Anthropic API whenever `ANTHROPIC_API_KEY` is set; mirror `scripts/tiling-check.py:351-352` `--allow-remote-ollama` precedent to fix (~1h). **Retrieval benchmark PASS**: 50 queries × 2 pipelines via `scripts/benchmark-runner.py`; v1.7 top-1 54.0% vs v1.6 baseline 24.0% (+30pp); error-reduction +39.5% vs ≥30% gate. Per-category breakdown in audit §6.2. **Competitor recheck (24h after compass May 16)**: kepano +1.1k★ to 31.6k, Copilot CLI integration issue still stale 3mo (genuine moat for us), NotebookLM May 2026 shipped Video Overviews + 4-tile Studio (widens derivative-outputs gap — filed M13 for v2.0 derive spec). **7-axis #1 verdict**: YES on 4 axes (compounding wiki, multi-writer safety, retrieval architecture free-tier, license openness), TIE on methodology (v1.8 closes), NO on GUI ergonomics (v2.5+) + derivative outputs (v2.0). Honest answer: #1 today on power-user-control axes, not in mainstream adoption without v2.0+v2.5.
|
||||
|
||||
**For post-compact resumption**: read `docs/audits/v1.7.1-fixes-plan.md` (this is your roadmap — 6 commits, ~2.5h, sequenced top-to-bottom with file paths + exact edits + verification steps + commit messages per fix). The plan starts with the BLOCKER (Fix 1) + Data Privacy callout (Fix 6) bundled. Working tree state on resume: 4 untracked files (audit doc, fixes plan, `scripts/baseline-v16.py`, `scripts/benchmark-runner.py`); `96a5505` wiki auto-commit landed the benchmark corpus at `wiki/meta/retrieval-benchmark-v1.7.md`; `make test` is 7/7 green; `bash bin/setup-retrieve.sh --no-llm` is provisioned (chunks/, bm25/, embed-cache.json exist — gitignored). User wants to "proceed" with the fixes after compact; do NOT push or tag without explicit go.
|
||||
|
||||
**Session record** (full prose, ~600 lines) in personal vault: `~/Documents/Obsidian Vault/sessions/2026-05-17 claude-obsidian v1.7 audit + fixes plan.md`. Ingest-log entry prepended at top of `~/Documents/Obsidian Vault/log/ingest-log.md` per global save convention.
|
||||
|
||||
2026-05-17: **v1.7.0 "Compound Vault" refoundation shipped locally** (branch `v1.7.0-compound-vault`, NOT pushed). Four workstreams committed as 4 separate feat commits: §3.1 substrate hard-prefer on `kepano/obsidian-skills` (9c8e510), §3.2 default transport with new `wiki-cli` skill + `scripts/detect-transport.sh` (6c7671e), §3.3 hybrid retrieval pipeline as opt-in `wiki-retrieve` skill with 4 new scripts + 2 hermetic test suites (45a5bd3), §3.4 multi-writer safety closing the latent corruption bug from v1.6 via `scripts/wiki-lock.sh` (66c11f9). Cross-cutting commit pending: version bump 1.6.0→1.7.0, README/CLAUDE.md updates, CHANGELOG entry, new `docs/compound-vault-guide.md` omnibus, this hot.md update. `make test` runs 7 suites green (was 3) — zero ollama / network dependency preserved. Plan file at `~/.claude/plans/read-in-full-the-hidden-sun.md`. User-paused at "full on review on all work done"; no push or tag until explicit go.
|
||||
|
||||
2026-04-24 (late night): v1.6.0 public release notes shipped. `docs/releases/v1.6.0.md` (Karpathy-style, 346 lines) establishes the release-notes convention. Three original SVGs at `wiki/meta/dragonscale-{mechanism-overview,6-test-flow,frontier-graph}.svg` carry the visual load; Wikipedia dragon curve referenced by text link only (no binary vendoring). R4 codex verifier ACCEPT WITH FIXES, 3 wording fixes applied. User runs `gh release create v1.6.0 --notes-file docs/releases/v1.6.0.md` when ready. Commits `85515bb` (docs), plus wiki/meta/ auto-commits for SVGs.
|
||||
|
||||
2026-04-24 (night): DragonScale end-to-end validation pass. Six-test menu run via Teams orchestration (codex gpt-5.4 for M1 dry-run, M1 commit, M4 autoresearch; chair for ollama pull, M2 allocate, M3 full tiling). All six green. First real fold committed (`wiki/folds/fold-k3-from-2026-04-23-to-2026-04-24-n8.md`, 115 lines, 8 children). First real tiling report at `wiki/meta/tiling-report-2026-04-24.md` (0 errors, 15 review pairs). M2 counter advanced 2 to 3, `c-000002` reserved-unassigned. M4 autoresearch filed 3 new concept pages (`Persistent Wiki Artifact`, `Source-First Synthesis`, `Query-Time Retrieval`) extending `[[How does the LLM Wiki pattern work?]]` with Karpathy gist + RAG + MemGPT + Obsidian docs as sources. v1.6.0 validated.
|
||||
|
||||
2026-04-24 (evening): v1.6.0 closeout via Teams approach (chair-led, codex gpt-5.4 for sub-agents). 2 explorers (closeout gaps + doc surface). 6 bounded writes (non-overlapping scope): `docs/dragonscale-guide.md` (new, 563 lines), `wiki/meta/2026-04-24-v1.6.0-release-session.md` (new, 346 lines), `wiki/meta/boundary-frontier-2026-04-24.md` (first real M4 run artifact, new), `docs/install-guide.md` (1.5.0 to 1.6.0 + M4 callout + flat-extractive correction), `README.md` (parenthetical + guide link), `wiki/hot.md` (drift fixes). 1 adversarial verifier returned ACCEPT WITH FIXES; all 11 fixes applied in place. Docs commit `eb1562f`. `make test` green (74+ assertions). Still no git tags for v1.5.0 / v1.5.1 / v1.6.0. User requested gpt-5.5; API rejects it on this codex CLI; gpt-5.4 used throughout.
|
||||
|
||||
2026-04-24 (late): Phase 4 shipped. Mechanism 4 (boundary-first autoresearch) implemented as `scripts/boundary-score.py` with expanded test coverage. `/autoresearch` without a topic now offers frontier candidates (opt-in, agenda-control labeled). Cross-file status updated. Version bumped to 1.6.0 in `plugin.json` + `marketplace.json`; no git tag created locally (only pre-DragonScale tags `v1.1` - `v1.4.3` exist).
|
||||
|
||||
2026-04-24 (afternoon): Phase 3.6 hardening, five surgical fixes (tiling --report path confinement, rollout baseline, AGENTS.md consistency, wiki-ingest .raw contradiction, install-guide version). v1.5.1.
|
||||
|
||||
2026-04-24 (morning): Phase 3.5 hardening pass. Cross-phase audit resolved 10 hold-ship items. At that point Mechanism 4 was marked NOT IMPLEMENTED (later reversed in Phase 4 the same day). `bin/setup-dragonscale.sh` + tests + Makefile added, CHANGELOG created, versions synced to 1.5.0.
|
||||
|
||||
2026-04-23 (3): Phase 3 complete. Semantic tiling lint shipped as opt-in. `scripts/tiling-check.py` with flock-guarded atomic cache, localhost-locked OLLAMA_URL default, symlink rejection, model-drift invalidation, and banded thresholds (error>=0.90, review>=0.80, conservative seeds). 4 codex review rounds, 10/10 accept.
|
||||
|
||||
2026-04-23 (2): Phase 2 complete. Deterministic page addresses MVP via `scripts/allocate-address.sh` (flock-guarded, recovers counter from max observed). New frontmatter `address: c-NNNNNN`. `wiki-ingest` and `wiki-lint` updated with opt-in Address Assignment and Validation sections. 3 codex rounds, 8/8 accept.
|
||||
|
||||
2026-04-23 (1): Phase 0-1 complete. DragonScale Memory spec (`wiki/concepts/DragonScale Memory.md` v0.3) plus `skills/wiki-fold/` for Mechanism 1 (log rollups, dry-run verified). Survived multi-round codex review.
|
||||
|
||||
## Plugin State
|
||||
|
||||
- **Version**: 1.7.1 (audit-driven patch on top of Compound Vault; plugin.json + marketplace.json synced; local-only branch `v1.7.0-compound-vault`, no push, no tag)
|
||||
- **Install ID**: `claude-obsidian@ai-marketing-hub-claude-obsidian`
|
||||
- **Skills**: 13 (wiki, wiki-ingest, wiki-query, wiki-lint, wiki-fold, save, autoresearch, canvas, defuddle, obsidian-bases, obsidian-markdown, **wiki-cli (v1.7)**, **wiki-retrieve (v1.7, opt-in)**)
|
||||
- **Scripts (v1.6)**: `scripts/allocate-address.sh`, `scripts/tiling-check.py`, `scripts/boundary-score.py` (DragonScale; opt-in; feature-detected by skills)
|
||||
- **Scripts (v1.7 — new)**: `scripts/detect-transport.sh`, `scripts/contextual-prefix.py`, `scripts/bm25-index.py`, `scripts/rerank.py`, `scripts/retrieve.py`, `scripts/wiki-lock.sh`
|
||||
- **Setup**: `bin/setup-vault.sh` (base vault), `bin/setup-dragonscale.sh` (opt-in DragonScale), `bin/setup-multi-agent.sh` (multi-agent bootstrap), `bin/setup-retrieve.sh` (opt-in v1.7 hybrid retrieval)
|
||||
- **Tests**: `make test` runs 7 suites — `test_allocate_address.sh`, `test_tiling_check.py`, `test_boundary_score.py`, **`test_bm25_index.py`**, **`test_retrieve.py`**, **`test_wiki_lock.sh`**, **`test_concurrent_write.sh`**. Zero ollama and zero network dependency for all core tests.
|
||||
- **Hooks**: 4 (SessionStart, PostCompact, PostToolUse [stages wiki/, .raw/, .vault-meta/; **v1.7: defers `git add` if wiki-lock locks held**], Stop)
|
||||
|
||||
## DragonScale Mechanisms
|
||||
|
||||
1. **Fold operator** (Mechanism 1): `skills/wiki-fold/`, dry-run verified AND first real fold committed at `wiki/folds/fold-k3-from-2026-04-23-to-2026-04-24-n8.md`.
|
||||
2. **Deterministic addresses** (Mechanism 2): shipped and exercised; vault counter at 3. `c-000001` on DragonScale Memory.md. `c-000002` reserved-unassigned from validation pass (gap acceptable per spec).
|
||||
3. **Semantic tiling lint** (Mechanism 3): shipped and activated. `nomic-embed-text` pulled; first tiling report at `wiki/meta/tiling-report-2026-04-24.md` (0 errors, 15 review-band pairs).
|
||||
4. **Boundary-first autoresearch** (Mechanism 4): shipped (Phase 4, opt-in). `scripts/boundary-score.py` + `tests/test_boundary_score.py`. `/autoresearch` without a topic surfaces top-5 frontier pages as candidates; user picks, overrides, or declines. Explicitly labeled "agenda control" in both spec and skill.
|
||||
|
||||
## Key Lessons from This Release Cycle
|
||||
|
||||
1. Cross-phase audits are essential. Individual phase reviews miss drift between phases.
|
||||
2. Opt-in feature detection (`[ -x script ] && [ -f state ]`) preserves default plugin behavior for adopters and non-adopters alike.
|
||||
3. PostToolUse hook matcher is `Write|Edit`, so Bash writes don't fire it. Scripts that mutate tracked state must be Bash-only to avoid side-effect commits.
|
||||
4. Seed-vault self-consistency matters: if the spec says post-rollout pages need addresses, the concept page itself has to have one.
|
||||
5. Codex adversarial review rounds stop when the punch list is empty, not when the author feels done.
|
||||
|
||||
## Style Preferences
|
||||
|
||||
- No em dashes (U+2014) or `--` as punctuation. Periods, commas, colons, or parentheses. Hyphens in compound words are fine.
|
||||
- Short and direct responses. No trailing summaries.
|
||||
- Parallel tool calls when independent.
|
||||
|
||||
## Active Threads
|
||||
|
||||
- DragonScale Mechanism 4 shipped in Phase 4 as an opt-in Topic Selection mode in `skills/autoresearch/`. All four DragonScale mechanisms are now shipped and feature-gated.
|
||||
- v1.6.0 not yet pushed to GitHub (local commits only, no git tag created). User controls push and tag timing.
|
||||
- CLAUDE.md has one pre-existing uncommitted change ("Release Blog Post" section) that predates this session.
|
||||
|
||||
## Repo Locations
|
||||
|
||||
- Working: `~/Desktop/claude-obsidian/`
|
||||
- Public: https://github.com/AI-Marketing-Hub/claude-obsidian
|
||||
@@ -0,0 +1,93 @@
|
||||
---
|
||||
type: meta
|
||||
title: "Wiki Index"
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- meta
|
||||
- index
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[overview]]"
|
||||
- "[[log]]"
|
||||
- "[[hot]]"
|
||||
- "[[dashboard]]"
|
||||
- "[[Wiki Map]]"
|
||||
- "[[concepts/_index]]"
|
||||
- "[[entities/_index]]"
|
||||
- "[[sources/_index]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Hot Cache]]"
|
||||
- "[[Compounding Knowledge]]"
|
||||
- "[[Andrej Karpathy]]"
|
||||
---
|
||||
|
||||
# Wiki Index
|
||||
|
||||
Last updated: 2026-04-15 | Total pages: 34 | Sources ingested: 2
|
||||
|
||||
Navigation: [[overview]] | [[log]] | [[hot]] | [[dashboard]] | [[Wiki Map]] | [[getting-started]]
|
||||
|
||||
---
|
||||
|
||||
## Concepts
|
||||
|
||||
- [[LLM Wiki Pattern]] — the pattern for building persistent, compounding knowledge bases using LLMs (status: mature)
|
||||
- [[Hot Cache]] — ~500-word session context file, updated after every ingest and session (status: mature)
|
||||
- [[Compounding Knowledge]] — why wiki knowledge grows more valuable over time, unlike RAG (status: mature)
|
||||
- [[cherry-picks]] — prioritized feature backlog from ecosystem research; 13 features to add to claude-obsidian (status: current)
|
||||
- [[SVG Diagram Style Guide]] — canonical visual style for all diagrams: Space Grotesk, #0A0A0A dark theme, #E07850 accent, full design tokens (status: evergreen)
|
||||
- [[Pro Hub Challenge]] — community challenge pattern for building claude-seo/claude-blog extensions; first challenge produced 6 submissions, 5 integrated in v1.9.0 (status: evergreen)
|
||||
- [[Semantic Topic Clustering]] — SERP-based keyword grouping replacing paid tools; hub-spoke architecture with interactive visualization (status: evergreen)
|
||||
- [[Search Experience Optimization]] — "read SERPs backwards" methodology for page-type mismatch detection and persona scoring (status: evergreen)
|
||||
- [[SEO Drift Monitoring]] — "git for SEO" baseline/diff/track with 17 comparison rules and SQLite persistence (status: evergreen)
|
||||
- [[DragonScale Memory]] — memory-layer spec inspired by the Heighway dragon curve; fold operator, deterministic page addresses, semantic tiling, boundary-first autoresearch (status: shipped v0.4, all four mechanisms opt-in)
|
||||
- [[Persistent Wiki Artifact]]: durable Markdown page as the LLM's memory object, distinct from ephemeral chat turns (status: developing)
|
||||
- [[Source-First Synthesis]]: provenance discipline; raw sources stay immutable while the wiki layer is synthesized and cited (status: developing)
|
||||
- [[Query-Time Retrieval]]: wiki query path synthesizes with citations; complementary to Obsidian's in-vault search (status: developing)
|
||||
|
||||
---
|
||||
|
||||
## Entities
|
||||
|
||||
- [[Andrej Karpathy]] — AI researcher, creator of the LLM Wiki pattern, former Tesla AI director (status: developing)
|
||||
- [[Ar9av-obsidian-wiki]] — multi-agent compatible LLM Wiki plugin; delta tracking manifest (status: current)
|
||||
- [[Nexus-claudesidian-mcp]] — native Obsidian plugin + MCP bridge; workspace memory, task management (status: current)
|
||||
- [[ballred-obsidian-claude-pkm]] — goal cascade PKM; auto-commit hooks, /adopt command (status: current)
|
||||
- [[rvk7895-llm-knowledge-bases]] — 3-depth query system, Marp slides, parallel deep research (status: current)
|
||||
- [[kepano-obsidian-skills]] — official skills from Obsidian creator; defuddle, obsidian-bases (status: current)
|
||||
- [[Claudian-YishenTu]] — native Obsidian plugin embedding Claude Code; plan mode, @mention (status: current)
|
||||
- [[Claude SEO]] — Tier 4 Claude Code skill for SEO analysis; 23 skills, 17 agents, 30 scripts at v1.9.0 (status: evergreen)
|
||||
|
||||
---
|
||||
|
||||
## Sources
|
||||
|
||||
- [[claude-obsidian-ecosystem-research]] — 2026-04-08 | web research across 16+ repos | 8 wiki pages created
|
||||
|
||||
---
|
||||
|
||||
## Questions
|
||||
|
||||
- [[How does the LLM Wiki pattern work]] — how the pattern works and why it outperforms RAG at human scale (status: developing)
|
||||
|
||||
---
|
||||
|
||||
## Comparisons
|
||||
|
||||
- [[Wiki vs RAG]] — when to use a wiki knowledge base versus RAG; verdict: wiki wins at <1000 pages
|
||||
- [[claude-obsidian-ecosystem]] — feature matrix of 16+ Claude+Obsidian projects; where claude-obsidian wins and gaps
|
||||
|
||||
---
|
||||
|
||||
## Decisions
|
||||
|
||||
- [[2026-04-14-community-cta-rollout]] - Skool community CTA footer added to 6 skill repos with per-tool frequency rules (status: active)
|
||||
- [[2026-04-15-slides-and-release-session]] - Claude SEO v1.9.0 slides (15-slide HTML deck) + GitHub release v1.9.0 with PDF asset (status: complete)
|
||||
- [[2026-04-15-release-report-session]] - Claude SEO v1.9.0 Release Report PDF: dark theme, 13 pages, WeasyPrint layout fixes, Challenge v2 added (status: complete)
|
||||
- [[2026-04-14-claude-seo-v190-session]] - Claude SEO v1.9.0 Pro Hub Challenge integration: 5 submissions, 4 new skills, 4 review rounds, cybersecurity audit (status: complete)
|
||||
|
||||
---
|
||||
|
||||
## Domains
|
||||
|
||||
<!-- Add domain entries here after scaffold -->
|
||||
+218
@@ -0,0 +1,218 @@
|
||||
---
|
||||
type: meta
|
||||
title: "Operation Log"
|
||||
updated: 2026-04-08
|
||||
tags:
|
||||
- meta
|
||||
- log
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[index]]"
|
||||
- "[[hot]]"
|
||||
- "[[overview]]"
|
||||
- "[[sources/_index]]"
|
||||
---
|
||||
|
||||
# Operation Log
|
||||
|
||||
Navigation: [[index]] | [[hot]] | [[overview]]
|
||||
|
||||
Append-only. New entries go at the TOP. Never edit past entries.
|
||||
|
||||
Entry format: `## [YYYY-MM-DD] operation | Title`
|
||||
|
||||
Parse recent entries: `grep "^## \[" wiki/log.md | head -10`
|
||||
|
||||
---
|
||||
|
||||
## [2026-04-24] save | v1.6.0 public release notes (Teams, Karpathy-style)
|
||||
- Type: release doc + visual assets
|
||||
- Locations (new): `docs/releases/v1.6.0.md` (346 lines, 6 sections, Karpathy-style prose), `wiki/meta/dragonscale-mechanism-overview.svg` (4-mechanism diagram with shared .vault-meta/ gate), `wiki/meta/dragonscale-6-test-flow.svg` (validation timeline), `wiki/meta/dragonscale-frontier-graph.svg` (M4 candidate + 3 filed pages)
|
||||
- Locations (modified): `wiki/meta/2026-04-24-v1.6.0-release-session.md` (cross-reference added pointing to public release notes)
|
||||
- Scope: Teams approach. R1 (chair) wrote 3 original SVGs per SVG Diagram Style Guide. R2 (codex worker) drafted Karpathy-style release prose. R3 (chair) stitched SVGs, pivoted Wikipedia imagery to text links only (no binary vendoring per permission). R4 (codex verifier) returned ACCEPT WITH FIXES, 3 wording fixes on version narrative. R5 (chair) applied fixes, committed.
|
||||
- Style: direct, short, signal-dense, lists over prose, no em dashes, no marketing terms. Verifier confirmed zero em-dashes and zero banned marketing language ('revolutionary', 'seamless', 'world-class', 'game-changing', 'unlock', 'transform').
|
||||
- Distribution (all three destinations covered): (1) `docs/releases/v1.6.0.md` public-facing file (commit `85515bb`), (2) `wiki/meta/2026-04-24-v1.6.0-release-session.md` internal engineering record (cross-linked), (3) GitHub Release body (user to paste from docs/releases/v1.6.0.md when ready to `gh release create v1.6.0`).
|
||||
- Wikipedia imagery: referenced as text link to `https://en.wikipedia.org/wiki/Dragon_curve` rather than hotlinked or vendored. Cleaner license-wise (no CC-BY-SA attribution needed) and no external dependency. The 3 original SVGs carry the visual load instead.
|
||||
- PII scan post-write: `docs/releases/v1.6.0.md` + all three SVGs are clean. No `/home/` paths, no real emails, no tokens.
|
||||
- Next recommended: user runs `gh release create v1.6.0 --notes-file docs/releases/v1.6.0.md` when ready to cut the public release. This also creates the annotated tag.
|
||||
|
||||
## [2026-04-24] save | DragonScale end-to-end validation pass (Teams, 6 tests)
|
||||
- Type: validation + first real fold + first real autoresearch
|
||||
- Tests executed (all green):
|
||||
- T0 ollama pull `nomic-embed-text`: done (274MB, 15s wall)
|
||||
- T1 M1 dry-run k=3 via codex: DRY-RUN OK, 8 children, no em-dashes
|
||||
- T2 M2 real allocate: counter advanced 2 to 3, got `c-000002` (unassigned reservation; gap acceptable per spec)
|
||||
- T3 M3 full tiling with model present: 41 pages scanned, 21 embedded, 20 correctly skipped (meta/excluded/embed-error), 0 errors at >=0.9, 15 pairs in 0.8-0.9 review band (top 0.8822 Compounding Knowledge vs LLM Wiki Pattern, a legitimate semantic neighbor), report at `wiki/meta/tiling-report-2026-04-24.md`
|
||||
- T4 M1 commit via codex: first real fold committed, `wiki/folds/fold-k3-from-2026-04-23-to-2026-04-24-n8.md` (115 lines, 8 children, flat extractive). Flips the long-standing "no fold committed yet" status
|
||||
- T6 M4 autoresearch no-topic via codex: selected "How does the LLM Wiki pattern work?" as candidate (score 1.7022, #3 after skipping top-1 source + top-2 self-reference); 6 web fetches (Karpathy gist, RAG paper arXiv 2005.11401, MemGPT arXiv 2310.08560, Obsidian docs); 3 new concept pages filed, each with Primary Sources
|
||||
- Locations (new): `wiki/folds/fold-k3-from-2026-04-23-to-2026-04-24-n8.md`, `wiki/meta/tiling-report-2026-04-24.md`, `wiki/concepts/Persistent Wiki Artifact.md`, `wiki/concepts/Source-First Synthesis.md`, `wiki/concepts/Query-Time Retrieval.md`
|
||||
- Locations (modified): `.vault-meta/address-counter.txt` (2 to 3), `wiki/index.md` (3 concept links), `wiki/concepts/_index.md` (3 concept links)
|
||||
- Scope: six-test menu the user approved. Codex gpt-5.4 for T1/T4/T6 (sub-agent delegation); chair for T0/T2/T3 (one-shot shell) and all integration (index, log, hot, commit).
|
||||
- Style: all new content uses colons or parens instead of em-dashes. Pre-existing em-dashes in index entries and wiki/concepts/_index.md left as-is (clean-room boundary; deferred to F-slice style pass).
|
||||
- Tests still green: `make test` passes (74+ assertions).
|
||||
- Integration: chair added the 3 new concepts to `wiki/index.md` and `wiki/concepts/_index.md` with colon-style descriptions so the fresh pages are discoverable. The cluster extends `[[How does the LLM Wiki pattern work?]]` and cross-references `[[LLM Wiki Pattern]]`.
|
||||
- Next recommended slice: either (G) commit this test batch and declare v1.6.0 validated, or (H) run a second fold k=3 now that 8 newer entries exist above this one and close the hierarchical-fold-not-yet-supported loop in a future phase.
|
||||
|
||||
## [2026-04-24] save | v1.6.0 closeout (Teams, chair-led)
|
||||
- Type: docs + release hygiene
|
||||
- Locations (new): wiki/meta/2026-04-24-v1.6.0-release-session.md (release session summary, 346 lines), wiki/meta/boundary-frontier-2026-04-24.md (first M4 run artifact against this vault), docs/dragonscale-guide.md (user-facing DragonScale guide, 563 lines)
|
||||
- Locations (modified): wiki/hot.md (tag-claim fix, Scripts line adds boundary-score, tests line adds test_boundary_score, push-line drift, tiling line-count, one em-dash), docs/install-guide.md (version 1.5.0 to 1.6.0, DragonScale callout expanded to all four mechanisms, "hierarchical log folds" corrected to "flat extractive log folds", points to docs/dragonscale-guide.md), README.md (DragonScale parenthetical expanded to all four mechanisms plus guide link)
|
||||
- Scope: Teams approach, chair-led. Slice A (2 codex read-only explorers: closeout punch list + doc-surface map). Slice B (6 bounded writes: 4 chair, 2 codex workers, non-overlapping write scopes). Slice C (codex adversarial verifier, ACCEPT WITH FIXES). Slice D (fix pass + log entry + manual commit of docs + README).
|
||||
- Verifier: C1 found 11 items across 6 files. All 11 applied. Flag typos `--allow-remote-ollama` and `--report PATH` corrected in release-session; boundary-frontier provenance corrected to `--top 7` to match default vs explicit top; hot.md tiling line-count claim stripped to avoid drift; hot.md "local tag only" corrected to "local commits only, no git tag"; install-guide log-fold wording corrected from "hierarchical" to "flat extractive"; dragonscale-guide rollback wording corrected (`.vault-meta/` is a shared gate across M2+M3+M4, not per-mechanism).
|
||||
- Model: codex gpt-5.4 used throughout. User requested gpt-5.5; not reachable via codex CLI 0.123.0 / this account at the time. models_cache lists max gpt-5.4, and the API rejects gpt-5.5 with "does not exist or you do not have access". Existing config already has `service_tier = "fast"` and `sandbox_mode = "workspace-write"`, matching the "fast for chatgpt with permission of full access" intent.
|
||||
- Tests: `make test` passes. test_allocate_address.sh (shell, 12 assertions), test_tiling_check.py (python, 18 assertions), test_boundary_score.py (python, 44 assertions). Zero ollama dependency.
|
||||
- Tags: still no local v1.5.0 / v1.5.1 / v1.6.0 tags. User controls tag creation and push. Pre-existing tags unchanged (v1.1, v1.4.0 through v1.4.3).
|
||||
- Deliberately NOT done: no real M1 fold committed; no M3 end-to-end run (needs `ollama pull nomic-embed-text`); pre-existing em-dashes in install-guide.md and README.md left untouched (clean-room boundary, not in write scope this slice); CLAUDE.md pre-existing uncommitted change left untouched.
|
||||
- Next recommended slice: either (E) push to origin/main and create annotated tags v1.5.0, v1.5.1, v1.6.0 in landing order, or (F) dedicated style pass to scrub pre-existing em-dashes across install-guide.md, README.md, and any other wiki files flagged by a grep scan.
|
||||
|
||||
## [2026-04-24] save | DragonScale Phase 4 — boundary-first autoresearch shipped (v1.6.0)
|
||||
- Type: feature release
|
||||
- Locations (new): scripts/boundary-score.py (with --top, --page, --json, stdout-only CLI), tests/test_boundary_score.py (40+ assertions)
|
||||
- Locations (modified): skills/autoresearch/SKILL.md (new Topic Selection section A/B/C with helper-failure fallback), commands/autoresearch.md (no-topic candidate flow with agenda-control label), wiki/concepts/DragonScale Memory.md (v0.4: M4 flipped from NOT IMPLEMENTED to shipped; exact formula without recency floor; filename-stem disclosure; fence-handling qualifiers), CHANGELOG.md, .claude-plugin/{plugin,marketplace}.json (1.5.0 -> 1.6.0), Makefile (test-boundary target), wiki/hot.md, wiki/index.md, wiki/concepts/_index.md (status drift resolved).
|
||||
- Scope: boundary-first autoresearch as opt-in Topic Selection mode. `/autoresearch` without a topic surfaces top-5 frontier pages; user picks/overrides/declines. Explicit helper-failure fallback to user-ask. Labeled "agenda control" throughout to match the spec's scope disclosure.
|
||||
- Correctness: filename-stem resolution including folder-qualified `[[notes/Foo]]` -> Foo.md. Self-loops, unresolved targets, meta-targets, symlinks, and vault escapes all excluded. Code-fence parser handles backticks AND tildes with CommonMark length tracking (longer opening fence is not closed by shorter inner fence). Indented blocks intentionally not filtered (Obsidian bullet convention).
|
||||
- Recency: exp(-days/30), no floor. Stale pages approach zero weight so they do not dominate frontier ranking.
|
||||
- Review rounds: codex adversarial Phase 4 round 1 (10 items: 7 reject + 3 refine). Round 2 (7 accept + 3 still-reject: folder-qualified stem, docstring floor mention, hot.md historical drift). Round 3 (3 accept, PASS).
|
||||
- Phase 3.6 (pre-Phase-4 hardening) already landed as v1.5.1: tiling --report VAULT_ROOT confinement, rollout baseline, AGENTS.md consistency, wiki-ingest .raw/ contradiction, install-guide version.
|
||||
- All four DragonScale mechanisms now shipped and opt-in. 44 commits ahead of origin/main, no push.
|
||||
|
||||
## [2026-04-24] save | DragonScale Phase 3.5 — cross-phase hardening to v1.5.0
|
||||
- Type: release hardening
|
||||
- Locations (new): bin/setup-dragonscale.sh (opt-in installer), tests/test_allocate_address.sh, tests/test_tiling_check.py, Makefile, CHANGELOG.md
|
||||
- Locations (modified): hooks/hooks.json (+.vault-meta/ staging), agents/wiki-ingest.md (single-writer rule for addresses), agents/wiki-lint.md (Mechanism 2+3 checks), skills/wiki-ingest/SKILL.md (aligned non-DragonScale wording), wiki/concepts/DragonScale Memory.md (M2 severity matches lint, M4 marked NOT IMPLEMENTED, seed page gets address c-000001), .claude-plugin/{plugin.json,marketplace.json} (1.4.2/1.4.3 → 1.5.0), README.md (11 skills + DragonScale callout), wiki/hot.md (refreshed for v1.5.0), .raw/.manifest.json (address_map now has DragonScale Memory.md → c-000001), .gitignore (.vault-meta/.tiling.lock + cache), .vault-meta/address-counter.txt (advanced to 2).
|
||||
- Scope: resolve the 10 hold-ship items from the cross-phase audit. Add reproducible test harness (make test passes). Version-bump plugin.json and marketplace.json to 1.5.0. Create CHANGELOG.md. Refresh hot cache.
|
||||
- Review rounds: codex 3.5a (5/5 accept on doc/agent fixes), codex final holistic (10/10 accept on audit items + 2 surgical regression fixes: wiki-ingest/wiki-lint non-DragonScale wording alignment, README skill count).
|
||||
- Tests: `make test` runs 12 shell assertions (allocator) + 18 python assertions (tiling-check). All pass; no ollama dependency.
|
||||
- Phase 3.5 complete. Repo state: 6 developer commits added this pass (f2e73c1, 2b49a0c, 8b28e48, 19ad7e4, 365f557, 2e7dd16). Total 39 commits ahead of origin/main. No push.
|
||||
|
||||
## [2026-04-24] save | DragonScale Phase 3 — semantic tiling MVP
|
||||
- Type: skill update + new script + threshold state
|
||||
- Locations: scripts/tiling-check.py (485 lines), .vault-meta/tiling-thresholds.json (seed defaults), skills/wiki-lint/SKILL.md (109-line Semantic Tiling section + item #10 in checks), wiki/concepts/DragonScale Memory.md (Mechanism 3 cost framing clarified)
|
||||
- Scope: opt-in embedding-based duplicate detection via ollama nomic-embed-text. Default bands error>=0.90, review>=0.80, explicitly documented as conservative seeds (not literature-backed interpolation). Calibration procedure documented, not automated.
|
||||
- Security: default OLLAMA_URL locked to 127.0.0.1; non-localhost requires --allow-remote-ollama flag. Symlinks and vault-root escapes rejected before file reads (prevents data exfil).
|
||||
- Correctness: cache keyed on sha256(model+body); orphan GC on save; model-drift auto-invalidation on load.
|
||||
- Concurrency: flock(LOCK_EX) on .vault-meta/.tiling.lock; per-PID temp file for atomic writes.
|
||||
- Scale: warn >500 pages; hard-fail exit 4 at >5000 pages.
|
||||
- Exit codes: 0/2/3/4/10/11 distinctly surfaced in wiki-lint wiring (not collapsed into "unknown").
|
||||
- Review rounds: 4 codex exec adversarial passes covering security, cache correctness, feature gate, inclusion logic, scale, threshold honesty, concurrency, exit codes, model drift, terminology coupling.
|
||||
Round 1: 10 items -> 7 reject + 3 refine.
|
||||
Round 2: 6 accept + 4 still-reject (symlink ordering, prose sync, exit-code wiring, terminology in checklist + "no API cost" claim).
|
||||
Round 3: 3 accept + 1 still-reject (cost-framing phrasing).
|
||||
Round 4: accept.
|
||||
- Final verdict: 10/10 accept.
|
||||
- Phase 3 complete. All three DragonScale mechanisms that were in-scope for the initial spec are now shipped as opt-in features. Mechanism 4 (boundary-first autoresearch) was flagged as agenda-control out-of-scope per the v0.2 scope boundary; may or may not ship as a future phase.
|
||||
|
||||
## [2026-04-23] save | DragonScale Phase 2 — deterministic page addresses MVP
|
||||
- Type: skill update + new script
|
||||
- Locations: scripts/allocate-address.sh, skills/wiki-ingest/SKILL.md (Address Assignment section), skills/wiki-lint/SKILL.md (Address Validation section), wiki/concepts/DragonScale Memory.md (Mechanism 2 rewritten v0.2→v0.3), .vault-meta/address-counter.txt, .raw/.manifest.json (new)
|
||||
- Scope: MVP address format `c-NNNNNN` (creation-order counter, zero-padded 6 digits). Rollout baseline 2026-04-23. Legacy pages exempt until deliberate backfill (future `l-` prefix). No content hash, no fold-ancestry encoding in the MVP (both deferred).
|
||||
- Concurrency: atomic allocation via flock-guarded Bash helper. Counter recovery from max observed `c-` address, never silent reset to 1.
|
||||
- Lint: post-rollout pages without address are errors; legacy pages without address are informational. Optional `.vault-meta/legacy-pages.txt` manifest grandfathers pages with missing/wrong `created:` metadata.
|
||||
- Re-ingest idempotency: `.raw/.manifest.json` `address_map` preserves path→address mapping across re-ingests and renames.
|
||||
- Naming: mechanism renamed from "content-addressable paths" to "deterministic page addresses" (the MVP is a counter, not a content hash; the old name was overclaim).
|
||||
- Review rounds: 2 codex exec adversarial passes. Round 1: 8 rejects covering counter mutation, race conditions, uniqueness atomicity, missing-file recovery, terminology drift, silent regression path, legacy classification, re-ingest idempotency. Round 2: 7 accept + 1 reject (manifest.json absent). Round 3 (item 8 only): accept after creating `.raw/.manifest.json`.
|
||||
- Final verdict: 8/8 accept.
|
||||
- Phase 2 complete. Phase 3 (semantic tiling lint) gated on human approval.
|
||||
|
||||
## [2026-04-23] save | DragonScale Phase 1 — wiki-fold skill shipped
|
||||
- Type: skill
|
||||
- Location: skills/wiki-fold/SKILL.md, skills/wiki-fold/references/fold-template.md
|
||||
- Scope: flat extractive fold over raw wiki/log.md entries. Dry-run default via Bash stdout (no Write tool, avoids PostToolUse hook residue). Structural idempotency via deterministic fold_id. Duplicate-range detection. Fold-of-folds explicitly out of scope.
|
||||
- Review rounds: 3 codex exec adversarial passes. Round 1: 1 refine + 6 reject across 7 items (allowed-tools, hook-mutation risk, idempotency claim, dry-run faithfulness, children structure, Mechanism 1 coverage, auto-commit conflict). Round 2: 6 accept + 1 reject (25/26 count inversion). Round 3 (item 4 only): accept.
|
||||
- Final verdict: 7/7 accept.
|
||||
- Dry-run artifact: /tmp/wiki-fold-dry-run-v2.md (not committed). fold_id: fold-k3-from-2026-04-10-to-2026-04-23-n8.
|
||||
- Phase 1 complete. Phase 2 (content-addressable paths) gated on human approval.
|
||||
|
||||
## [2026-04-23] save | DragonScale Memory v0.2 — post-adversarial-review
|
||||
- Type: concept revision
|
||||
- Location: wiki/concepts/DragonScale Memory.md
|
||||
- Review: codex exec adversarial review rejected all 7 load-bearing claims in v0.1
|
||||
- Changes: weakened LSM analogy, removed strong prompt-cache claim, replaced 0.85 threshold with calibration procedure, justified 2^k as MVP convenience, acknowledged scope-boundary leak for boundary-first autoresearch, added Operational Policies section (retention/tombstones/versioning/conflict/concurrency/provenance/ACL), tagged claims as [sourced]/[derived]/[conjecture], narrowed tagging scope per re-review
|
||||
- Re-review result: 7/7 accepted (after one surgical fix on tagging-scope language)
|
||||
- Phase 0 complete. Phase 1 (wiki-fold skill) gated on human approval.
|
||||
|
||||
## [2026-04-23] save | DragonScale Memory — Phase 0 design doc (proposed)
|
||||
- Type: concept
|
||||
- Location: wiki/concepts/DragonScale Memory.md
|
||||
- From: brainstorming session on applying Heighway dragon curve properties to LLM wiki memory architecture
|
||||
- Scope: memory-layer only, NOT agent reasoning. Four mechanisms: (1) fold operator (LSM-style exponential compaction at 2^k log entries), (2) content-addressable page paths for prompt-cache stability, (3) semantic tiling lint (embedding-based dedup, 0.85 cosine threshold), (4) boundary-first autoresearch scoring
|
||||
- Status: proposed. Phase 0 pending codex adversarial review. Phase 1+ (fold skill, address anchors, tiling lint, boundary score) gated on review pass.
|
||||
- Primary sources verified: Dragon curve (Wikipedia, boundary dim 1.523627086), Regular paperfolding sequence (OEIS A014577), LSM trees (arXiv 2504.17178, LevelDB 10x level ratio), MemGPT (arXiv 2310.08560), Anthropic prompt caching docs (5min/1hr TTL, 20-block lookback)
|
||||
- Links updated: wiki/concepts/_index.md, wiki/index.md
|
||||
|
||||
## [2026-04-15] save | Claude SEO v1.9.0 Slides and GitHub Release
|
||||
- Type: session
|
||||
- Location: wiki/meta/2026-04-15-slides-and-release-session.md
|
||||
- From: built 15-slide HTML presentation deck (v190.html), fixed hardcoded path in release_report.py, pushed 68 files to GitHub, tagged v1.9.0, created GitHub release with PDF asset
|
||||
- Key lessons: Path.home() not hardcoded paths, git pull --rebase before big pushes, Chrome blocks file:// cross-origin images, .claude/ always in .gitignore
|
||||
- Release: https://github.com/AgriciDaniel/claude-seo/releases/tag/v1.9.0
|
||||
|
||||
## [2026-04-15] save | Claude SEO v1.9.0 Release Report — PDF Complete
|
||||
- Type: session
|
||||
- Location: wiki/meta/2026-04-15-release-report-session.md
|
||||
- From: full session completing the v1.9.0 PDF release report. Dark theme, 13 pages, 1.53 MB. Fixed logo (double-space filename), empty spaces, page-break orphans, file:// URL encoding.
|
||||
- Key fixes: `urllib.parse.quote()` for file:// URIs; `display:table-cell` is atomic in WeasyPrint (no page-break); fixed `height:297mm` causes empty space; replaced orphan tables with paragraphs
|
||||
- Challenge v2 added: keyword LEADS, $600 prize pool, deadline April 28
|
||||
- Output: `~/Desktop/Claude-SEO-v1.9.0-Release-Report.pdf`
|
||||
|
||||
## [2026-04-14] save | Claude SEO v1.9.0 — Pro Hub Challenge Integration Session
|
||||
- Type: session + 4 concept pages + 1 entity page
|
||||
- Location: wiki/meta/2026-04-14-claude-seo-v190-session.md
|
||||
- From: full v1.9.0 implementation session — reviewed 5 community submissions, integrated 4 new skills (seo-cluster, seo-sxo, seo-drift, seo-ecommerce), enhanced seo-hreflang, added DataForSEO cost guardrails
|
||||
- Pages created: [[2026-04-14-claude-seo-v190-session]], [[Claude SEO]], [[Pro Hub Challenge]], [[Semantic Topic Clustering]], [[Search Experience Optimization]], [[SEO Drift Monitoring]]
|
||||
- Review rounds: 4 (code review x3 + cybersecurity audit). Score: 87 → 93 → 97 → 85 security
|
||||
- Key learnings: always verify subagent output (40-line count error caught), insertion-point bugs caught by max-effort plan review, pre-existing security debt identified (10 of 15 findings)
|
||||
|
||||
## [2026-04-14] save | SVG Diagram Style Guide
|
||||
- Type: concept
|
||||
- Location: wiki/concepts/SVG Diagram Style Guide.md
|
||||
- From: extracted design tokens from 17 production SVGs in claude-ads/assets/diagrams/
|
||||
- Covers: colors, typography, layout primitives, card patterns, arrow connectors, numbered circles, file naming
|
||||
|
||||
## [2026-04-14] save | Community CTA Footer Rollout
|
||||
- Type: decision
|
||||
- Location: wiki/meta/2026-04-14-community-cta-rollout.md
|
||||
- From: session adding Skool community footer to 6 skill repos (claude-ads, claude-seo, claude-obsidian, claude-blog, banana-claude, claude-cybersecurity)
|
||||
- Key insight: frequency calibration per tool type; single-point orchestrator instruction pattern
|
||||
|
||||
## [2026-04-10] save | Backlink Empire - Blog Posts, Karpathy Gist, GitHub Cross-Linking
|
||||
- Type: session
|
||||
- Location: wiki/meta/2026-04-10-backlink-empire-session.md
|
||||
- From: full session covering blog creation (claude-obsidian + claude-canvas), Karpathy gist comment, 26 GitHub README updates with Author/community/backlink sections, homepage URLs on 10 repos, topics on 25 repos, rankenstein.pro backlinks on 5 SEO repos
|
||||
- Blog posts: agricidaniel.com/blog/claude-obsidian-ai-second-brain, agricidaniel.com/blog/claude-canvas-ai-visual-production
|
||||
- Impact: ~87 new backlinks from DA 96 github.com, 6 rankenstein.pro backlinks, 25 Skool community links
|
||||
|
||||
## [2026-04-08] save | claude-obsidian v1.4 Release Session
|
||||
- Type: session
|
||||
- Location: wiki/meta/claude-obsidian-v1.4-release-session.md
|
||||
- From: full release cycle covering v1.1 (URL/vision/delta tracking, 3 new skills), v1.4.0 (audit response, multi-agent compat, Bases dashboard, em dash scrub, security history rewrite), and v1.4.1 (plugin install command hotfix)
|
||||
- Key lessons: plugin install is 2-step (marketplace add then install), allowed-tools is not valid frontmatter, Bases uses filters/views/formulas not Dataview syntax, hook context does not survive compaction, git filter-repo needs 2 passes for full scrub
|
||||
|
||||
## [2026-04-08] ingest | Claude + Obsidian Ecosystem Research
|
||||
- Type: research ingest
|
||||
- Source: `.raw/claude-obsidian-ecosystem-research.md`
|
||||
- Queries: 6 parallel web searches + 12 repo deep-reads
|
||||
- Pages created: [[claude-obsidian-ecosystem]], [[cherry-picks]], [[claude-obsidian-ecosystem-research]], [[Ar9av-obsidian-wiki]], [[Nexus-claudesidian-mcp]], [[ballred-obsidian-claude-pkm]], [[rvk7895-llm-knowledge-bases]], [[kepano-obsidian-skills]], [[Claudian-YishenTu]]
|
||||
- Key finding: 16+ active Claude+Obsidian projects; 13 cherry-pick features identified for v1.3.0+
|
||||
- Top gap confirmed: no delta tracking, no URL ingestion, no auto-commit
|
||||
|
||||
## [2026-04-07] session | Full Audit, System Setup & Plugin Installation
|
||||
- Type: session
|
||||
- Location: wiki/meta/full-audit-and-system-setup-session.md
|
||||
- From: 12-area repo audit, 3 fixes, plugin installed to local system, folder renamed
|
||||
|
||||
## [2026-04-07] session | claude-obsidian v1.2.0 Release Session
|
||||
- Type: session
|
||||
- Location: wiki/meta/claude-obsidian-v1.2.0-release-session.md
|
||||
- From: full build session — v1.2.0 plan execution, cosmic-brain→claude-obsidian rename, legal/security audit, branded GIFs, PDF install guide, dual GitHub repos
|
||||
|
||||
|
||||
- Source: `.raw/` (first ingest)
|
||||
- Pages updated: [[index]], [[log]], [[hot]], [[overview]]
|
||||
- Key insight: The wiki pattern turns ephemeral AI chat into compounding knowledge — one user dropped token usage by 95%.
|
||||
|
||||
## [2026-04-07] setup | Vault initialized
|
||||
|
||||
- Plugin: claude-obsidian v1.1.0
|
||||
- Structure: seed files + first ingest complete
|
||||
- Skills: wiki, wiki-ingest, wiki-query, wiki-lint, save, autoresearch
|
||||
@@ -0,0 +1,95 @@
|
||||
---
|
||||
type: session
|
||||
title: "Backlink Empire - Blog Posts, Karpathy Gist, GitHub Cross-Linking"
|
||||
created: 2026-04-10
|
||||
updated: 2026-04-10
|
||||
tags:
|
||||
- session
|
||||
- backlinks
|
||||
- seo
|
||||
- github
|
||||
- blog
|
||||
- rankenstein
|
||||
status: complete
|
||||
related:
|
||||
- "[[Claude Obsidian]]"
|
||||
- "[[Claude Canvas]]"
|
||||
- "[[Rankenstein]]"
|
||||
- "[[Karpathy LLM Wiki Pattern]]"
|
||||
decision_date: 2026-04-10
|
||||
---
|
||||
|
||||
## What Was Done
|
||||
|
||||
### Blog Posts Created
|
||||
|
||||
Two blog posts written, deployed to Vercel, and submitted to Google Indexing API + Bing IndexNow:
|
||||
|
||||
1. **claude-obsidian-ai-second-brain** - "I Turned Obsidian Into a Self-Organizing AI Brain"
|
||||
- Focus keyword: "obsidian ai second brain"
|
||||
- ~2,800 words, 5 repo images/GIFs, 3 SVG charts, 8 sourced statistics
|
||||
- Images from `wiki/meta/` in the claude-obsidian repo (graph view, wiki map, welcome canvas GIF)
|
||||
- Live: agricidaniel.com/blog/claude-obsidian-ai-second-brain
|
||||
|
||||
2. **claude-canvas-ai-visual-production** - "Claude Code Just Turned Obsidian Canvas Into an AI Design Studio"
|
||||
- Focus keyword: "obsidian canvas ai"
|
||||
- ~2,500 words, 5 screenshots from repo, 2 SVG charts, 7 sourced statistics
|
||||
- Images from `assets/screenshots/` in the claude-canvas repo
|
||||
- Live: agricidaniel.com/blog/claude-canvas-ai-visual-production
|
||||
|
||||
### Karpathy Gist Comment
|
||||
|
||||
Comment posted on Andrej Karpathy's LLM Wiki gist (gist ID: 442a6bf555914893e9891c11519de94f). Links claude-obsidian (358 stars), claude-canvas, and the blog post. Highlights hot cache, contradiction flagging, 8-category lint, and autonomous research loops as differentiators.
|
||||
|
||||
### GitHub Backlink Empire (26 Repos Updated)
|
||||
|
||||
**Phase 1 - API updates (no README changes):**
|
||||
- Set homepage URLs on 10 repos (on-page-seo, Keywordo-kun, claude-youtube, marketing-skill-pack, google-maps-scraper, claude-repurpose, claude-gif, claude-avatar, rankenstein-pro-latest, claude-canvas)
|
||||
- Set topics/tags on 25 repos (grouped by category: SEO, content, marketing, obsidian, media, dev tools, n8n)
|
||||
|
||||
**Phase 2-5 - README updates:**
|
||||
- Added standardized Author section to 25 repos with links to:
|
||||
- agricidaniel.com/about
|
||||
- agricidaniel.com/blog
|
||||
- skool.com/ai-marketing-hub
|
||||
- youtube.com/@AgriciDaniel
|
||||
- github.com/AgriciDaniel
|
||||
|
||||
**Rankenstein.pro backlinks (5 SEO-relevant repos only):**
|
||||
- claude-seo: "Publishing Pipeline" section
|
||||
- claude-blog: "Publishing Platform" section
|
||||
- on-page-seo: blockquote before Author
|
||||
- Keywordo-kun: blockquote before Author
|
||||
- marketing-skill-pack: mentioned in Author context
|
||||
|
||||
### Verification Results
|
||||
|
||||
64/65 checks passed. 1 minor note: claude-obsidian uses footer-style attribution instead of formal `## Author` heading (intentional - already had custom attribution).
|
||||
|
||||
## Key Decisions
|
||||
|
||||
- **rankenstein.pro placement**: Only on SEO-relevant repos (5 of 26). Not spammed across video/image/dev tools. Keeps it natural.
|
||||
- **Karpathy gist tone**: Technical, value-first. Led with the implementation, not marketing. Matched the collaborative tone of the comment thread.
|
||||
- **Blog keyword strategy**: "obsidian ai second brain" (trending, moderate competition) and "obsidian canvas ai" (low competition, growing query). Both exploit competitor gaps - no existing article covers a specific tool in depth.
|
||||
- **Cover images**: Used real repo assets (pixel-art covers, screenshots, GIFs) rather than stock photos. Converts better and is more authentic.
|
||||
|
||||
## Numbers
|
||||
|
||||
- ~87 new backlinks from github.com (DA 96) to agricidaniel.com
|
||||
- ~6 backlinks to rankenstein.pro
|
||||
- ~25 backlinks to skool.com/ai-marketing-hub
|
||||
- 22 total pages in sitemap (was 20 at session start)
|
||||
- 15 total blog posts (was 13)
|
||||
|
||||
## Workflow for Future Blog Posts
|
||||
|
||||
This session established a repeatable workflow:
|
||||
1. Explore repo thoroughly (features, images, stats)
|
||||
2. Keyword research + competitor analysis via web search
|
||||
3. Write blog JSON with HTML content, SVG charts, repo images
|
||||
4. Add to blogPosts.ts, sitemap.xml, llms.txt
|
||||
5. Build + prerender + deploy to Vercel
|
||||
6. Submit to Google Indexing API + Bing IndexNow
|
||||
7. Set repo homepage to blog URL
|
||||
8. Add blog backlink to repo README
|
||||
9. Update README with Author/community section
|
||||
@@ -0,0 +1,77 @@
|
||||
---
|
||||
type: meta
|
||||
title: "Claude SEO v1.9.0 — Pro Hub Challenge Integration Session"
|
||||
created: 2026-04-14
|
||||
updated: 2026-04-14
|
||||
tags:
|
||||
- session
|
||||
- claude-seo
|
||||
- v1.9.0
|
||||
- pro-hub-challenge
|
||||
- release
|
||||
status: complete
|
||||
related:
|
||||
- "[[Claude SEO]]"
|
||||
- "[[Pro Hub Challenge]]"
|
||||
- "[[Semantic Topic Clustering]]"
|
||||
- "[[Search Experience Optimization]]"
|
||||
- "[[SEO Drift Monitoring]]"
|
||||
- "[[E-commerce SEO]]"
|
||||
---
|
||||
|
||||
# Claude SEO v1.9.0 — Pro Hub Challenge Integration
|
||||
|
||||
**Date**: 2026-04-14
|
||||
**Duration**: Extended session (~4 hours)
|
||||
**Scope**: Integrate 5 community submissions from the AI Marketing Hub Pro Hub Challenge into claude-seo
|
||||
|
||||
## What Was Done
|
||||
|
||||
### Community Submissions Integrated
|
||||
| Contributor | Submission | Integrated As |
|
||||
|------------|------------|--------------|
|
||||
| **Lutfiya Miller** (Winner) | Semantic Cluster Engine | `seo-cluster` — SERP overlap clustering, hub-spoke architecture, interactive visualization |
|
||||
| **Florian Schmitz** | SXO Skill | `seo-sxo` — page-type mismatch detection, SERP-to-user-story, persona scoring |
|
||||
| **Dan Colta** | SEO Drift Monitor | `seo-drift` — baseline/diff/track with 17 comparison rules, SQLite persistence |
|
||||
| **Chris Muller** | Multi-lingual SEO | `seo-hreflang` enhancements — cultural profiles (DACH, FR, ES, JA), locale formats, content parity audit |
|
||||
| **Matej Marjanovic** | E-commerce + DataForSEO Cost Config | `seo-ecommerce` + cost guardrails with approval workflow |
|
||||
|
||||
### Numbers
|
||||
- **Before**: 19 skills, 13 agents, 23 scripts (v1.8.2)
|
||||
- **After**: 23 skills, 17 agents, 30 scripts (v1.9.0)
|
||||
- **New files created**: 30
|
||||
- **Existing files modified**: 31
|
||||
- **Total lines added**: ~5,500
|
||||
|
||||
### Architecture Decisions
|
||||
1. **SEO parts only** — blog-specific features (translation, multilingual pipeline, character images) stay for claude-blog
|
||||
2. **Full integration with optional execution** — cluster skill outputs content briefs when claude-blog isn't detected, full execution when it is
|
||||
3. **Security-hardened drift scripts** — original had SSRF bypass via curl fallback. Completely rewritten using only fetch_page.py
|
||||
4. **Cost guardrails** — threshold-based approval, daily limits, file locking, audit trail on reset
|
||||
|
||||
## Review Process (4 rounds)
|
||||
|
||||
| Round | Type | Score | Issues Found |
|
||||
|-------|------|-------|-------------|
|
||||
| 1 | superpowers:code-reviewer (3 agents) | 87/100 | 6 critical (step numbering, SSRF fallback, install.ps1, counts, CHANGELOG, README) |
|
||||
| 2 | superpowers:code-reviewer (3 agents) | 89/100 → 93/100 after fixes | 8 important (drift history routing, marketplace.json, audit math, AGENTS.md, CONTRIBUTING) |
|
||||
| 3 | superpowers:code-reviewer (3 agents) | 97/100 | 5 suggestions only (all pre-existing) |
|
||||
| 4 | /cybersecurity (8 agents) | 77/100 → 85/100 after fixes | H3: cost bypass, M4: XSS, M3: CI, M5: locking, L5: pyproject |
|
||||
|
||||
### Security Findings & Fixes
|
||||
- **XSS in cluster-map.html** — `truncate()` wasn't wrapped in `escapeHtml()`. Fixed.
|
||||
- **Cost guardrail bypass** — `reset` + unknown endpoint = unlimited spend. Fixed: reset requires `--confirm` + audit trail, unknown endpoints return `needs_approval`.
|
||||
- **File locking** — cost ledger had no locking for parallel agents. Fixed with fcntl.
|
||||
- **Pre-existing (deferred)**: validate_url DNS rebinding, install script injection, OAuth file permissions, no pip lockfile
|
||||
|
||||
## Key Learnings
|
||||
|
||||
1. **Agent output verification is essential** — subagents got seo/SKILL.md line count wrong by 40 lines, miscounted skills (25 vs 23), and would have placed CONTRIBUTING.md section in wrong location (orphaning subsections)
|
||||
2. **Security audits find real bugs** — the XSS and cost guardrail bypass were genuine issues that static review missed
|
||||
3. **Pre-existing vs new** — of 15 security findings, only 5 were introduced by v1.9.0. The codebase has technical debt from earlier versions
|
||||
4. **Plan review catches insertion-point bugs** — the max-effort plan review found 2 bugs (CONTRIBUTING.md section placement, README command ordering) before they were executed
|
||||
|
||||
## Files for Reference
|
||||
- Plan: `~/.claude/plans/smooth-popping-pebble.md`
|
||||
- CHANGELOG: v1.9.0 entry in `~/Desktop/Claude-SEO/CHANGELOG.md`
|
||||
- Contributors: `~/Desktop/Claude-SEO/CONTRIBUTORS.md`
|
||||
@@ -0,0 +1,58 @@
|
||||
---
|
||||
type: decision
|
||||
title: "Community CTA Footer Rollout"
|
||||
created: 2026-04-14
|
||||
updated: 2026-04-14
|
||||
decision_date: 2026-04-14
|
||||
status: active
|
||||
tags:
|
||||
- marketing
|
||||
- skool
|
||||
- community
|
||||
- growth
|
||||
related:
|
||||
- "[[index]]"
|
||||
---
|
||||
|
||||
# Community CTA Footer Rollout
|
||||
|
||||
AI Marketing Hub Skool community links (free + pro) added as a footer to 6 Claude Code skill repos. The footer appears after major deliverables, never during mid-workflow or on quick utilities.
|
||||
|
||||
## The Footer
|
||||
|
||||
```
|
||||
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
|
||||
Built by agricidaniel - Join the AI Marketing Hub community
|
||||
Free -> https://www.skool.com/ai-marketing-hub
|
||||
Pro -> https://www.skool.com/ai-marketing-hub-pro
|
||||
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
|
||||
```
|
||||
|
||||
## Implementation Pattern
|
||||
|
||||
Single-point instruction in each repo's orchestrator SKILL.md. One section controls the footer text, show list, and skip list. No duplication across sub-skills.
|
||||
|
||||
## Per-Repo Frequency Calibration
|
||||
|
||||
| Repo | Triggers | Rationale |
|
||||
|------|----------|-----------|
|
||||
| claude-ads | 12 commands | Audits, reports, analyses (each is a session-level event) |
|
||||
| claude-seo | 12 commands | Audits, technical, content (same pattern as ads) |
|
||||
| claude-obsidian | 3 operations | Only scaffold, lint, autoresearch (high-frequency tool, conservative) |
|
||||
| claude-blog | 8 commands | Write, rewrite, audit, analyze, brief, strategy, calendar, geo. Explicit guard: never in generated blog content/HTML |
|
||||
| banana-claude | 4 commands | Image generate, edit, batch (skip chat, inspire, config) |
|
||||
| claude-cybersecurity | All audits | Single-purpose tool, every completed report gets it |
|
||||
|
||||
## Design Principles
|
||||
|
||||
1. Free link listed first. Pro framed as "support the creator," not a gate.
|
||||
2. Footer appears only after value is delivered, never before or during.
|
||||
3. High-frequency tools (obsidian, banana chat) get fewer triggers to avoid spam.
|
||||
4. Content-producing tools (blog) explicitly exclude CTA from generated output.
|
||||
5. Single source of truth per repo. Update one section to change everything.
|
||||
|
||||
## Future Considerations
|
||||
|
||||
- Add "once per conversation" cap if power users report repetition across back-to-back commands.
|
||||
- Track conversion rate. If zero joins after months, experiment with wording.
|
||||
- Forks will strip the CTA. That is fine and expected under MIT license.
|
||||
@@ -0,0 +1,61 @@
|
||||
---
|
||||
type: meta
|
||||
title: "Claude SEO v1.9.0 Release Report Session"
|
||||
updated: 2026-04-15
|
||||
tags:
|
||||
- meta
|
||||
- session
|
||||
- claude-seo
|
||||
- pdf
|
||||
- weasyprint
|
||||
status: complete
|
||||
related:
|
||||
- "[[Claude SEO]]"
|
||||
- "[[Pro Hub Challenge]]"
|
||||
- "[[2026-04-14-claude-seo-v190-session]]"
|
||||
---
|
||||
|
||||
# Claude SEO v1.9.0 Release Report Session
|
||||
|
||||
Date: 2026-04-15 | Output: `~/Desktop/Claude-SEO-v1.9.0-Release-Report.pdf`
|
||||
|
||||
## What Was Built
|
||||
|
||||
13-page dark-theme PDF release report for Claude SEO v1.9.0. Generated via `scripts/release_report.py` using WeasyPrint + matplotlib. Covers: Pro Hub Challenge contributions, architecture evolution, review score progression, security audit findings, DataForSEO cost guardrails, and Challenge v2 announcement.
|
||||
|
||||
**Stats**: 13 pages, 1.53 MB, 4 charts, 7 screenshots embedded, logo visible on title page.
|
||||
|
||||
**Brand**: Space Grotesk font, `#0A0A0A` background, `#E07850` accent (rust-orange), `#111111` cards, `#2D2D2D` borders. Matches SVG Diagram Style Guide.
|
||||
|
||||
## Bugs Fixed
|
||||
|
||||
| Bug | Root Cause | Fix |
|
||||
|-----|-----------|-----|
|
||||
| Logo not rendering | Filename has double space: "AI MArketing hub pro logo with white text.png" | Corrected path in `generate_report()` |
|
||||
| `file://` images not loading | Spaces in paths not URL-encoded | Added `_file_url()` helper with `urllib.parse.quote()` |
|
||||
| Review checker false WARNs | Checked URL-encoded paths against filesystem | `unquote()` before `Path.exists()` |
|
||||
| Title page empty bottom half | Fixed `height:297mm` + sparse content | Removed fixed height, added "In This Report" card |
|
||||
| Contributor card page-break orphans | `display:table-cell` is atomic in WeasyPrint | Replaced two-column layout with stacked blocks |
|
||||
| Architecture scripts table orphaned | 7-row table split across pages | Replaced table with paragraph |
|
||||
| Security highlight box orphaned | Orphaned after large table | Merged text into intro paragraph |
|
||||
| Chart page mostly empty | Chart too small relative to page | Increased figsize height, moved chart first in section |
|
||||
|
||||
## WeasyPrint PDF Lessons
|
||||
|
||||
1. **`file://` URIs must be URL-encoded** — spaces become `%20`. Use `urllib.parse.quote(path, safe="/:@")`. When reviewing paths extracted from HTML, use `unquote()` before `Path.exists()`.
|
||||
2. **`display:table-cell` is atomic** — WeasyPrint cannot break a table-cell across pages. For content that might span pages (contributor cards, multi-row content), use stacked block elements (`<p>` + `<ul>`) instead of two-column table layouts.
|
||||
3. **Fixed height causes empty space** — `height: 297mm` on a div with sparse content leaves blank below. Use auto height + generous padding instead.
|
||||
4. **Tables that overflow: replace with paragraphs** — if a table consistently orphans its last rows, a `<p>` with inline `<code>` spans is more reliable.
|
||||
5. **Chart figsize controls page fill** — matplotlib figsize directly affects how much of the page the chart occupies. Increase height to fill empty space after a chart.
|
||||
6. **`max-height: 165mm` on `.chart-container img`** — good default for charts on their own section page.
|
||||
7. **Check filenames carefully** — "AI MArketing hub pro logo with white text.png" has a double space between "hub" and "pro". `Path.exists()` is the fastest way to catch this.
|
||||
|
||||
## Pro Hub Challenge v2 (April)
|
||||
|
||||
Added to the report's "What's Next" section. Details:
|
||||
- Keyword: **LEADS**
|
||||
- Prizes: $400 (1st) + $200 (2nd) in Claude Credits
|
||||
- Deadline: April 28, 2026
|
||||
- Scope: anything touching lead generation — Claude Code skills, n8n workflows, MCP servers, scrapers, dashboards, pipelines
|
||||
- Rules: GitHub repo or .zip + 1-2 min demo video, must be functional, solo or team welcome
|
||||
- Previous winner: Lutfiya Miller (seo-cluster, integrated in v1.9.0)
|
||||
@@ -0,0 +1,79 @@
|
||||
---
|
||||
type: meta
|
||||
title: "Claude SEO v1.9.0 Slides and GitHub Release Session"
|
||||
updated: 2026-04-15
|
||||
tags:
|
||||
- meta
|
||||
- session
|
||||
- claude-seo
|
||||
- github
|
||||
- slides
|
||||
status: complete
|
||||
related:
|
||||
- "[[Claude SEO]]"
|
||||
- "[[Pro Hub Challenge]]"
|
||||
- "[[2026-04-14-claude-seo-v190-session]]"
|
||||
- "[[2026-04-15-release-report-session]]"
|
||||
---
|
||||
|
||||
# Claude SEO v1.9.0 Slides and GitHub Release Session
|
||||
|
||||
Date: 2026-04-15 | Outputs: `claude-seo-slides/v190.html`, GitHub release v1.9.0
|
||||
|
||||
## What Was Built
|
||||
|
||||
### HTML Slide Deck (claude-seo-slides/v190.html)
|
||||
|
||||
15-slide community presentation for the v1.9.0 release. Scroll-snap HTML, no external library. Matches the existing v1.7.2 dark-theme brand exactly.
|
||||
|
||||
**Tech pattern:**
|
||||
- `scroll-snap-type: y mandatory` on `html`, each slide `min-height:100vh` + `scroll-snap-align: start`
|
||||
- `IntersectionObserver` per slide to update progress bar and nav dots
|
||||
- Keyboard: ArrowDown/Right/Space to advance, ArrowUp/Left to go back
|
||||
- `file:///` absolute paths for local screenshots with `onerror` fallback handlers
|
||||
|
||||
**Brand:** `#0A0A0A` bg, `#E07850` coral accent, Space Grotesk headings, IBM Plex Mono body. `.claude/`, `.superpowers/` added to `.gitignore` before push.
|
||||
|
||||
**Slide structure (15 slides):**
|
||||
|
||||
| # | Title | Key Content |
|
||||
|---|-------|-------------|
|
||||
| 01 | Title | 23 skills, 5 contributors, 4 new skills, 30 scripts |
|
||||
| 02 | Executive Summary | 8 metric cards, community wins, technical wins |
|
||||
| 03 | The Challenge | 3 cards, full 8-stage timeline table |
|
||||
| 04 | Community Posts | Announcement + winner screenshots (local file paths) |
|
||||
| 05 | Contributors | All 6, with Winner/Proficient/Reviewed badges |
|
||||
| 06 | seo-cluster | Lutfiya Miller, features, screenshot, integration notes |
|
||||
| 07 | seo-sxo | Florian Schmitz, detection example, screenshot |
|
||||
| 08 | seo-drift | Dan Colta, flow diagram, features, screenshot |
|
||||
| 09 | seo-ecommerce | Matej Marjanovic, cost approval box, screenshot |
|
||||
| 10 | seo-hreflang | Chris Muller, cultural profiles table, screenshot |
|
||||
| 11 | Architecture Evolution | Before/after counts, 7 new scripts list |
|
||||
| 12 | Review Process | Score timeline 87→93→97→85, findings table per round |
|
||||
| 13 | Security Audit | 85/100, detailed fixes table |
|
||||
| 14 | DataForSEO Guardrails | Bypass chain, before/after code snippet, fcntl |
|
||||
| 15 | What's Next | v1.9.1 H1/H2/M1 deferred items, Challenge v2 LEADS |
|
||||
|
||||
**Screenshot paths note:** `claude-seo-slides/v190.html` contains 7 absolute `file://` home paths for community post screenshots. Not sensitive, but not portable. `onerror` handlers show placeholder text when images fail. Works in Firefox; Chrome blocks cross-origin `file://` image requests.
|
||||
|
||||
### GitHub Release v1.9.0
|
||||
|
||||
**Steps taken:**
|
||||
1. Fixed `SCREENSHOTS_DIR` hardcoded path in `scripts/release_report.py`: replaced the old absolute home Downloads path with `Path.home() / "Downloads" / "..."` (Path was already imported).
|
||||
2. Added `.claude/` and `.superpowers/` to `.gitignore`.
|
||||
3. Staged 68 files (31 modified, 37 new), committed as `feat: v1.9.0 Pro Hub Challenge community integration`.
|
||||
4. Remote had 1 commit ahead ("Remove blog links from README") — resolved with `git pull --rebase`.
|
||||
5. Tagged `v1.9.0` on HEAD, pushed tag.
|
||||
6. Created GitHub release via `gh release create v1.9.0` with PDF attached (`Claude-SEO-v1.9.0-Release-Report.pdf`). No HTML slides attached as release asset.
|
||||
|
||||
**Release URL:** https://github.com/AgriciDaniel/claude-seo/releases/tag/v1.9.0
|
||||
|
||||
**Commit stats:** 68 files, 9,662 insertions, 51 deletions.
|
||||
|
||||
## Key Lessons
|
||||
|
||||
1. **`Path.home()` for user-relative paths in scripts** — never hardcode `/home/username/...`. Use `Path.home() / "..."` or `os.path.expanduser("~")`. Catches before push with a simple `grep -rn "/home/"`.
|
||||
2. **Always `git pull --rebase` before pushing a big local commit** — even on solo repos with active GitHub Actions or web edits. Avoids a merge commit cluttering the history.
|
||||
3. **`gh release create` attaches assets directly** — pass file path as positional argument. Only attach what users actually need to download (PDF), not presentation assets (HTML) that live in the repo.
|
||||
4. **`.claude/` and `.superpowers/` should always be in `.gitignore`** — they hold project-specific Claude Code permissions and plugin state. Not credentials, but not repo content either.
|
||||
5. **Chrome blocks `file://` cross-origin image requests** — HTML files opened as `file://` cannot load images from other `file://` paths in Chrome. Firefox allows it. For portable local HTML with images, use `python3 -m http.server` or embed images as base64 data URIs.
|
||||
@@ -0,0 +1,348 @@
|
||||
---
|
||||
type: meta
|
||||
title: "2026-04-24 v1.6.0 Release Session"
|
||||
created: 2026-04-24
|
||||
updated: 2026-04-24
|
||||
tags:
|
||||
- meta
|
||||
- session
|
||||
- release
|
||||
- dragonscale
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[log]]"
|
||||
- "[[hot]]"
|
||||
- "[[DragonScale Memory]]"
|
||||
- "[[claude-obsidian-v1.4-release-session]]"
|
||||
---
|
||||
|
||||
# 2026-04-24 v1.6.0 Release Session
|
||||
|
||||
A same-day release cycle covering v1.5.0, v1.5.1, and v1.6.0.
|
||||
|
||||
This session took DragonScale from a three-mechanism opt-in extension to a four-mechanism opt-in extension, with a hardening point release in the middle.
|
||||
|
||||
**Public release notes**: [docs/releases/v1.6.0.md](../../docs/releases/v1.6.0.md). That file is the user-facing artifact and the source to paste into a `gh release create v1.6.0` body. This session page is the internal engineering record.
|
||||
|
||||
Ground truth for this page comes from:
|
||||
|
||||
- `CHANGELOG.md` entries `[1.5.0]`, `[1.5.1]`, and `[1.6.0]`
|
||||
- `wiki/log.md` entries for 2026-04-24 covering Phase 3.5, Phase 3.6, and Phase 4
|
||||
- current repository state for file existence, tags, and commit history
|
||||
|
||||
## Release Sequence
|
||||
|
||||
- `v1.5.0`: Phase 3.5, morning. Cross-phase hardening, opt-in DragonScale installer, reproducible tests, `CHANGELOG.md`, `Makefile`, version sync.
|
||||
- `v1.5.1`: Phase 3.6, afternoon. Five surgical hardening fixes across tiling security, rollout baseline, AGENTS docs, wiki-ingest wording, and install guide versioning.
|
||||
- `v1.6.0`: Phase 4, late. DragonScale Mechanism 4 shipped as opt-in boundary-first autoresearch, with the new scorer script, new tests, and no-topic autoresearch candidate flow.
|
||||
|
||||
## Release Summary
|
||||
|
||||
The morning release, v1.5.0, was the hardening pass that closed the hold-ship audit items accumulated across Phases 1 through 3.
|
||||
|
||||
The afternoon release, v1.5.1, was a narrow correction pass that fixed five remaining inconsistencies and one tiling write-path confinement issue before Phase 4 moved forward.
|
||||
|
||||
The late release, v1.6.0, shipped the previously deferred fourth DragonScale mechanism and completed the opt-in DragonScale set described in `[[DragonScale Memory]]`.
|
||||
|
||||
No git tag was created for v1.5.0, v1.5.1, or v1.6.0.
|
||||
|
||||
No push to `origin/main` happened during this release cycle.
|
||||
|
||||
`git tag` verifies that only pre-DragonScale tags exist locally:
|
||||
|
||||
- `v1.1`
|
||||
- `v1.4.0`
|
||||
- `v1.4.1`
|
||||
- `v1.4.2`
|
||||
- `v1.4.3`
|
||||
|
||||
## Phase Timeline For 2026-04-24
|
||||
|
||||
### Morning: Phase 3.5 to v1.5.0
|
||||
|
||||
This was the cross-phase hardening release.
|
||||
|
||||
The goal was to close the 10 hold-ship items found by the holistic audit across the earlier DragonScale phases.
|
||||
|
||||
The release added the missing installer, test harness, and changelog artifacts that made the extension reproducible instead of partially documented.
|
||||
|
||||
Mechanism 4 remained intentionally unshipped at this point.
|
||||
|
||||
The spec still marked it as not implemented, and `/autoresearch` was unchanged.
|
||||
|
||||
The release also synchronized plugin and marketplace version strings to `1.5.0`.
|
||||
|
||||
### Afternoon: Phase 3.6 to v1.5.1
|
||||
|
||||
This was a targeted hardening point release.
|
||||
|
||||
The work did not expand scope.
|
||||
|
||||
It corrected five issues that mattered for correctness, consistency, or security:
|
||||
|
||||
- tiling report path confinement
|
||||
- rollout baseline date
|
||||
- AGENTS skill-table and compatibility wording
|
||||
- wiki-ingest wording about immutable `.raw/` versus maintained manifest
|
||||
- install guide version drift
|
||||
|
||||
The result was a smaller, cleaner release state before Mechanism 4 moved from proposal to shipped code.
|
||||
|
||||
### Late: Phase 4 to v1.6.0
|
||||
|
||||
This was the feature release that shipped boundary-first autoresearch as DragonScale Mechanism 4.
|
||||
|
||||
The new scorer computes frontier candidates from the wikilink graph and exposes them when `/autoresearch` is invoked without a topic in a DragonScale-enabled vault.
|
||||
|
||||
The mode remained opt-in and was labeled as agenda control throughout the implementation and documentation.
|
||||
|
||||
Versions were then synced to `1.6.0` in `plugin.json` and `marketplace.json`.
|
||||
|
||||
## What Shipped In v1.5.0
|
||||
|
||||
### Added
|
||||
|
||||
- `skills/wiki-fold/` as the shipped opt-in fold operator for Mechanism 1. The changelog describes it as extractive and structurally idempotent, with dry-run by default and explicit commit mode.
|
||||
- `scripts/allocate-address.sh` and the `address: c-NNNNNN` convention for Mechanism 2. The allocator is flock-guarded and supports re-ingest idempotency through `.raw/.manifest.json address_map`.
|
||||
- `scripts/tiling-check.py` for Mechanism 3. The checker uses local `nomic-embed-text` through ollama and exposes conservative seeded thresholds for duplicate-page review.
|
||||
- `wiki/concepts/DragonScale Memory.md` as the full design spec for the extension at v0.3 during this release.
|
||||
- `bin/setup-dragonscale.sh` as the idempotent DragonScale installer.
|
||||
- a reproducible `tests/` suite for allocator and tiling checks, runnable through `make test`
|
||||
- `Makefile` developer targets for testing, setup, and cleanup
|
||||
- `CHANGELOG.md` as the first explicit release ledger for this branch of work
|
||||
|
||||
### Changed
|
||||
|
||||
- `hooks/hooks.json` stages `.vault-meta/` in addition to `wiki/` and `.raw/`, so DragonScale runtime state is captured by the auto-commit hook
|
||||
- `skills/wiki-ingest/SKILL.md` and `skills/wiki-lint/SKILL.md` gained opt-in DragonScale sections behind feature detection, while preserving default behavior for vaults that have not adopted DragonScale
|
||||
- `agents/wiki-ingest.md` added the single-writer rule for address assignment
|
||||
- `agents/wiki-lint.md` documented Address Validation and Semantic Tiling checks
|
||||
- stale `allowed-tools` frontmatter was removed from `wiki-ingest` and `wiki-lint` to match the kepano-style `name` plus `description` convention for those files
|
||||
- version strings were synced across plugin metadata and documentation
|
||||
|
||||
### Security
|
||||
|
||||
`scripts/tiling-check.py` defaults `OLLAMA_URL` to localhost only.
|
||||
|
||||
Remote endpoints require the `--allow-remote-ollama` flag.
|
||||
|
||||
Symlinks and vault-root escapes are rejected before any read.
|
||||
|
||||
### Explicitly Not In v1.5.0
|
||||
|
||||
Mechanism 4 was still a future proposal in this release.
|
||||
|
||||
The changelog states that boundary-first autoresearch did not ship in v1.5.0 and that `skills/autoresearch/SKILL.md` was unchanged.
|
||||
|
||||
That matters because the same day later reversed this state in v1.6.0.
|
||||
|
||||
## What Shipped In v1.5.1
|
||||
|
||||
v1.5.1 was intentionally narrow.
|
||||
|
||||
It fixed five concrete issues and did not introduce a new mechanism.
|
||||
|
||||
### Fixed
|
||||
|
||||
- `scripts/tiling-check.py`: the `--report PATH` flag is now resolved against `VAULT_ROOT` and rejected if it escapes, preventing accidental or hostile writes outside the vault
|
||||
- `.vault-meta/legacy-pages.txt`: rollout baseline corrected from `2026-04-24` to `2026-04-23`, which matches the earliest addressed page in the seed vault
|
||||
- `AGENTS.md`: `wiki-fold` added to the skills table, and the blanket statement about all skills using only `name` and `description` was narrowed to newer skills only
|
||||
- `skills/wiki-ingest/SKILL.md`: clarified the difference between immutable user-dropped source documents in `.raw/` and the maintained `.raw/.manifest.json`
|
||||
- `docs/install-guide.md`: version updated from `1.2.0` to `1.5.0`, with a DragonScale optional-install callout
|
||||
|
||||
### Why v1.5.1 Existed
|
||||
|
||||
The log entry describes this as pre-Phase-4 hardening.
|
||||
|
||||
That is the right framing.
|
||||
|
||||
It reduced drift across docs, setup state, and security behavior before the Mechanism 4 rollout.
|
||||
|
||||
## What Shipped In v1.6.0
|
||||
|
||||
### Added
|
||||
|
||||
- boundary-first autoresearch as DragonScale Mechanism 4, still opt-in
|
||||
- `scripts/boundary-score.py`, which computes `(out_degree - in_degree) * recency_weight` over the wikilink graph
|
||||
- `/autoresearch` without a topic now offers the top five frontier pages as candidate research targets when DragonScale is present
|
||||
- `tests/test_boundary_score.py` with 35 unit tests covering parsing, recency weight, wikilink extraction, graph construction, scoring, and CLI behavior
|
||||
- `make test-boundary`, integrated into `make test`
|
||||
|
||||
### Changed
|
||||
|
||||
- `skills/autoresearch/SKILL.md` gained a Topic Selection section with explicit, boundary-first, and user-ask paths
|
||||
- `commands/autoresearch.md` documents no-topic usage for the two relevant modes
|
||||
- `wiki/concepts/DragonScale Memory.md` flipped Mechanism 4 from not implemented to shipped, added the exact scoring formula, and bumped the spec to v0.4
|
||||
- version metadata was synced to `1.6.0` across plugin metadata
|
||||
|
||||
### Scope And Behavior
|
||||
|
||||
The log entry is more detailed than the changelog and records the operating constraints clearly:
|
||||
|
||||
- the mode is opt-in
|
||||
- the mode is labeled as agenda control
|
||||
- helper failure falls back to asking the user
|
||||
- self-loops, unresolved targets, meta-targets, symlinks, and vault escapes are excluded
|
||||
- code-fence parsing handles both backticks and tildes with length-aware closure rules
|
||||
- stale pages approach zero recency weight instead of receiving a floor
|
||||
|
||||
### Why v1.6.0 Matters
|
||||
|
||||
This is the release that completed the originally proposed four-part DragonScale design in shipped form.
|
||||
|
||||
It also changed `/autoresearch` from topic-only operation to a dual-mode flow that can surface candidate frontiers while still keeping user choice explicit.
|
||||
|
||||
## DragonScale Mechanisms: Opt-In Status
|
||||
|
||||
All four mechanisms are opt-in.
|
||||
|
||||
No mechanism is always-on for base vault adopters.
|
||||
|
||||
1. Fold operator (Mechanism 1). Shipped as `skills/wiki-fold/`. Dry-run verified. No fold committed in this vault.
|
||||
2. Deterministic page addresses (Mechanism 2). Shipped via `scripts/allocate-address.sh` and `address: c-NNNNNN`. Feature-detected by ingest and lint flows.
|
||||
3. Semantic tiling lint (Mechanism 3). Shipped via `scripts/tiling-check.py`. Uses local ollama embedding flow and remains optional.
|
||||
4. Boundary-first autoresearch (Mechanism 4). Shipped in v1.6.0 via `scripts/boundary-score.py` and no-topic `/autoresearch` candidate selection.
|
||||
|
||||
## Verified New Files Mentioned In This Session
|
||||
|
||||
The following files were explicitly requested for verification and do exist in the repository:
|
||||
|
||||
- `scripts/boundary-score.py`: yes. New Mechanism 4 scorer in v1.6.0.
|
||||
- `tests/test_boundary_score.py`: yes. New Mechanism 4 unit tests in v1.6.0.
|
||||
- `bin/setup-dragonscale.sh`: yes. New opt-in installer in v1.5.0.
|
||||
- `CHANGELOG.md`: yes. Created in v1.5.0.
|
||||
- `Makefile`: yes. Created in v1.5.0 and later extended in v1.6.0.
|
||||
|
||||
Additional new files recorded in the 2026-04-24 log for Phase 3.5 also include:
|
||||
|
||||
- `tests/test_allocate_address.sh`
|
||||
- `tests/test_tiling_check.py`
|
||||
|
||||
## Testing And Validation Notes
|
||||
|
||||
The release log gives exact validation points that are worth preserving here.
|
||||
|
||||
For v1.5.0:
|
||||
|
||||
- `make test` ran 12 shell assertions for the allocator
|
||||
- `make test` ran 18 Python assertions for tiling-check
|
||||
- no ollama dependency was required for those core tests
|
||||
|
||||
For v1.6.0:
|
||||
|
||||
- `tests/test_boundary_score.py` added 35 unit tests
|
||||
- `make test-boundary` was added and folded into `make test`
|
||||
|
||||
The session also records adversarial review rounds for each phase:
|
||||
|
||||
- Phase 3 reached final verdict `10/10 accept`
|
||||
- Phase 3.5 recorded `5/5 accept` on the doc and agent fixes plus `10/10 accept` on the final holistic audit pass
|
||||
- Phase 4 reached pass after three adversarial rounds
|
||||
|
||||
Those review outcomes explain why the day was split into a hardening release, a point release, and then the final feature release.
|
||||
|
||||
## Git State And Release State
|
||||
|
||||
### Version State
|
||||
|
||||
Local metadata was bumped to `1.5.0`, then `1.5.1`, then `1.6.0` across the release cycle.
|
||||
|
||||
The user explicitly noted that v1.6.0 shipped locally and that no release tag exists for it.
|
||||
|
||||
Current repo checks support that statement.
|
||||
|
||||
### Tag State
|
||||
|
||||
`git tag` returns:
|
||||
|
||||
- `v1.1`
|
||||
- `v1.4.0`
|
||||
- `v1.4.1`
|
||||
- `v1.4.2`
|
||||
- `v1.4.3`
|
||||
|
||||
There is no local tag for `v1.5.0`, `v1.5.1`, or `v1.6.0`.
|
||||
|
||||
### Push State
|
||||
|
||||
The 2026-04-24 log entries repeatedly state `No push` or `no push`.
|
||||
|
||||
This release session remained local.
|
||||
|
||||
There was no push to `origin/main`.
|
||||
|
||||
### DragonScale Commit Range
|
||||
|
||||
The earliest DragonScale-labeled commit in repository history is:
|
||||
|
||||
- `16d9923` `feat: add flock-guarded address allocator for DragonScale Mechanism 2`
|
||||
|
||||
The latest commit in the release-phase DragonScale range before later wiki auto-commits is:
|
||||
|
||||
- `09399ae` `chore: bump to v1.6.0 with Phase 4 + Phase 3.6 CHANGELOG entries`
|
||||
|
||||
For the feature endpoint specifically, the latest DragonScale mechanism commit is:
|
||||
|
||||
- `ad1a1d0` `feat: autoresearch integrates DragonScale Mechanism 4 (opt-in)`
|
||||
|
||||
This page treats `16d9923` through `09399ae` as the practical release-session range, because the latter is the final version-sync commit that closed the feature release.
|
||||
|
||||
## Key Lessons
|
||||
|
||||
These are pulled from the `wiki/hot.md` Key Lessons section current at the time of this session.
|
||||
|
||||
1. Cross-phase audits are essential. Individual phase reviews miss drift between phases.
|
||||
2. Opt-in feature detection (`[ -x script ] && [ -f state ]`) preserves default plugin behavior for adopters and non-adopters alike.
|
||||
3. PostToolUse hook matcher is `Write|Edit`. Bash writes do not fire it. Scripts that mutate tracked state must be Bash-only to avoid side-effect commits.
|
||||
4. Seed-vault self-consistency matters. If the spec says post-rollout pages need addresses, the concept page itself has to have one.
|
||||
5. Codex adversarial review rounds stop when the punch list is empty, not when the author feels done.
|
||||
|
||||
## Session Shape Compared With v1.4
|
||||
|
||||
The earlier `[[claude-obsidian-v1.4-release-session]]` documented a broad ecosystem-response cycle that bundled research, audit response, style cleanup, security history rewrite, and a plugin install hotfix.
|
||||
|
||||
This 2026-04-24 session was narrower and more implementation-focused.
|
||||
|
||||
The main pattern was:
|
||||
|
||||
- complete the missing hardening work for existing DragonScale mechanisms
|
||||
- correct the remaining drift with a point release
|
||||
- then ship the previously deferred fourth mechanism
|
||||
|
||||
That sequencing was justified by the log and changelog.
|
||||
|
||||
Mechanism 4 was not simply appended to an unstable tree.
|
||||
|
||||
It was delayed until the earlier mechanism set had been hardened and the docs aligned.
|
||||
|
||||
## What This Session Did Not Do
|
||||
|
||||
It did not create release tags for the three versions shipped that day.
|
||||
|
||||
It did not push any of the DragonScale release work to `origin/main`.
|
||||
|
||||
It did not make DragonScale mandatory for users who only want the base wiki plugin.
|
||||
|
||||
It did not change the fact that all DragonScale mechanisms remain feature-detected and opt-in.
|
||||
|
||||
## Canonical Release Narrative
|
||||
|
||||
If this day needs to be summarized in one paragraph later, the clean version is this:
|
||||
|
||||
Morning v1.5.0 made DragonScale reproducible and internally consistent. Afternoon v1.5.1 cleaned up the remaining hardening issues. Late v1.6.0 shipped the previously deferred boundary-first autoresearch mechanism, completing the four-part DragonScale extension without changing the plugin's opt-in default posture.
|
||||
|
||||
## Verified Facts Worth Reusing
|
||||
|
||||
- `CHANGELOG.md` contains explicit entries for `[1.5.0]`, `[1.5.1]`, and `[1.6.0]`
|
||||
- `wiki/log.md` contains same-day entries for Phase 3.5, Phase 3.6, and Phase 4
|
||||
- `plugin.json` and `marketplace.json` are at `1.6.0`
|
||||
- the DragonScale installer exists as `bin/setup-dragonscale.sh`
|
||||
- the Mechanism 4 scorer exists as `scripts/boundary-score.py`
|
||||
- the Mechanism 4 tests exist as `tests/test_boundary_score.py`
|
||||
- `Makefile` and `CHANGELOG.md` exist
|
||||
- only pre-DragonScale tags are present locally
|
||||
- there was no push to `origin/main`
|
||||
|
||||
## Closing Note
|
||||
|
||||
The important distinction for future sessions is that `v1.5.0`, `v1.5.1`, and `v1.6.0` are real local release states with changelog and log backing, but they are not git-tagged releases in this repository snapshot.
|
||||
@@ -0,0 +1,67 @@
|
||||
---
|
||||
type: meta
|
||||
title: "Boundary Frontier Snapshot (2026-04-24)"
|
||||
updated: 2026-04-24
|
||||
tags:
|
||||
- meta
|
||||
- dragonscale
|
||||
- mechanism-4
|
||||
status: snapshot
|
||||
related:
|
||||
- "[[DragonScale Memory]]"
|
||||
- "[[log]]"
|
||||
- "[[hot]]"
|
||||
---
|
||||
|
||||
# Boundary Frontier Snapshot (2026-04-24)
|
||||
|
||||
Navigation: [[index]] | [[log]] | [[DragonScale Memory]]
|
||||
|
||||
First end-to-end run of DragonScale Mechanism 4 (`scripts/boundary-score.py`) against this vault. Generated from `./scripts/boundary-score.py --json --top 7` at 2026-04-24T08:49:16Z.
|
||||
|
||||
## What this is
|
||||
|
||||
This is a scoring snapshot, not a prescription. The boundary score heuristic surfaces pages that are outward-pointing and recently-touched as candidates for `/autoresearch` to extend. It is explicitly agenda-control per the [[DragonScale Memory]] spec, v0.4, Mechanism 4: the ranking shapes what the agent researches next, and a user should accept, override, or decline any candidate.
|
||||
|
||||
Formula: `boundary_score(p) = (out_degree(p) - in_degree(p)) * exp(-age_days / 30)`.
|
||||
|
||||
No recency floor. Pages older than ~90 days approach zero weight by design, so a stale hub does not dominate the frontier.
|
||||
|
||||
## Frontier (top 7, score > 0)
|
||||
|
||||
| # | score | out | in | age_d | title | path |
|
||||
|---|---|---|---|---|---|---|
|
||||
| 1 | 4.693 | 8 | 0 | 16 | Claude + Obsidian Ecosystem Research | wiki/sources/claude-obsidian-ecosystem-research.md |
|
||||
| 2 | 4.000 | 4 | 0 | 0 | DragonScale Memory | wiki/concepts/DragonScale Memory.md |
|
||||
| 3 | 1.702 | 3 | 0 | 17 | How does the LLM Wiki pattern work? | wiki/questions/How does the LLM Wiki pattern work.md |
|
||||
| 4 | 1.135 | 2 | 0 | 17 | Wiki vs RAG | wiki/comparisons/Wiki vs RAG.md |
|
||||
| 5 | 0.717 | 1 | 0 | 10 | SEO Drift Monitoring | wiki/concepts/SEO Drift Monitoring.md |
|
||||
| 6 | 0.717 | 1 | 0 | 10 | Search Experience Optimization (SXO) | wiki/concepts/Search Experience Optimization.md |
|
||||
| 7 | 0.717 | 1 | 0 | 10 | Semantic Topic Clustering | wiki/concepts/Semantic Topic Clustering.md |
|
||||
|
||||
22 scoreable pages total (meta, fold, and index pages excluded).
|
||||
|
||||
## Reading the result
|
||||
|
||||
- Row 1 is the ecosystem research source. It links out to eight entity pages and is not linked back, which is expected for a raw source: it seeds the graph rather than being referenced by it. The score is correct; following this candidate would extend one of its eight entities rather than re-examining the source itself.
|
||||
- Row 2 (DragonScale Memory) has age_days=0 and zero in-degree. This is a fresh concept page not yet linked back by any discussion. A legitimate frontier signal.
|
||||
- Rows 3-7 are older pages (~10 to 17 days) with modest out-degree. The recency decay correctly damps them relative to fresh pages.
|
||||
- No page ranks on pure recency with zero out-degree, because the formula multiplies degree-delta by recency.
|
||||
|
||||
## Calibration note
|
||||
|
||||
The halflife of 30 days was chosen as a default, not a tuned value. If this vault grows past ~100 pages and out-degree patterns change, the halflife should be reviewed alongside the weighting between degree and recency. The [[DragonScale Memory]] spec explicitly tags these as seed values, not literature-backed.
|
||||
|
||||
## Reproduce
|
||||
|
||||
```
|
||||
./scripts/boundary-score.py --json --top 7
|
||||
```
|
||||
|
||||
Read-only. Requires python3 only. No DragonScale setup needed to run the scorer itself.
|
||||
|
||||
## Connections
|
||||
|
||||
- [[DragonScale Memory]]: spec, Mechanism 4
|
||||
- [[log]]: operation log
|
||||
- [[hot]]: recent context
|
||||
Binary file not shown.
|
After Width: | Height: | Size: 403 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 363 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 847 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 1.3 MiB |
@@ -0,0 +1,116 @@
|
||||
---
|
||||
type: session
|
||||
title: "claude-obsidian v1.2.0 Release Session"
|
||||
created: 2026-04-07
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- session
|
||||
- release
|
||||
- plugin
|
||||
- github
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[getting-started]]"
|
||||
- "[[index]]"
|
||||
- "[[overview]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
---
|
||||
|
||||
# claude-obsidian v1.2.0 Release Session
|
||||
|
||||
Full build, audit, polish, and community release of the claude-obsidian plugin + vault kit. Previously named `cosmic-brain`.
|
||||
|
||||
---
|
||||
|
||||
## What Was Built
|
||||
|
||||
### Phase 1 — Critical Fixes
|
||||
- `marketplace.json`: version corrected `1.0.0→1.2.0`, owner metadata updated
|
||||
- `main.canvas`: removed 5 broken file node references (gitignored files that don't exist for community users)
|
||||
- `community-plugins.json`: deduplicated from 6→4 canonical entries: `[excalidraw, banners, calendar, thino]`
|
||||
|
||||
### Phase 2 — Vault Onboarding
|
||||
- `wiki/getting-started.md`: new onboarding page inside the vault (3-step quick start, hot cache explanation, command reference, navigation links)
|
||||
- `wiki/index.md`: Entities, Questions, Comparisons sections populated with existing seed pages
|
||||
- `wiki/meta/dashboard.md`: Dataview queries fixed — was querying `answer_quality` and `confidence` fields that don't exist in seed pages; replaced with `status` and `updated`
|
||||
- `CLAUDE.md`: placeholder text replaced with actual vault description
|
||||
- `wiki/canvases/welcome.canvas`: CTA node added pointing to getting-started and `/wiki`
|
||||
|
||||
### Phase 3 — README + Docs
|
||||
- README: complete pre-installed plugins table, CSS snippets section, Banner usage section, file structure updated
|
||||
- `bin/setup-vault.sh`: success message now lists all 4 pre-installed plugins and 3 CSS snippets
|
||||
|
||||
### Phase 4 — PDF Install Guide
|
||||
- `docs/install-guide.md`: full printable install guide (prerequisites, 3 install options, first steps, command reference, plugin guide, MCP setup, troubleshooting)
|
||||
- `docs/install-guide.pdf`: 159KB, generated via `npx md-to-pdf`
|
||||
|
||||
### Phase 5 — Version Bump
|
||||
- `plugin.json` and `marketplace.json` bumped to `1.2.0`
|
||||
|
||||
---
|
||||
|
||||
## Rename: cosmic-brain → claude-obsidian
|
||||
|
||||
Full project rename executed:
|
||||
- GitHub repos renamed: `AgriciDaniel/cosmic-brain` → `AI-Marketing-Hub/claude-obsidian` (public), `avalonreset-pro/cosmic-brain` → `avalonreset-pro/claude-obsidian` (private)
|
||||
- Local directory: `~/cosmic-brain/` → `~/claude-obsidian/`
|
||||
- All text references updated across 14 files via sed
|
||||
- `wiki/meta/cosmic-brain-cover.gif` renamed to `wiki/meta/claude-obsidian-cover.gif`
|
||||
|
||||
---
|
||||
|
||||
## Legal & Security
|
||||
|
||||
### Security Audit
|
||||
- No API keys, tokens, or secrets found in any tracked file
|
||||
- No private keys or certificates
|
||||
- All credential references in docs are placeholder values
|
||||
- Excalidraw `main.js` correctly NOT tracked despite audit agent claiming otherwise
|
||||
|
||||
### Legal Fixes
|
||||
- `LICENSE`: MIT license file created (was declared in plugin.json but file was missing)
|
||||
- `ITS-Dataview-Cards.css` + `ITS-Image-Adjustments.css`: GPL-2.0 attribution headers added
|
||||
|
||||
### .gitignore Tightened
|
||||
Added rules to prevent future accidental commits of: video files (`*.mkv`, `*.mp4`), transcripts, scratch canvases (`Untitled *.canvas`, `*Images.canvas`), and personal images in vault root.
|
||||
|
||||
---
|
||||
|
||||
## Visual / README
|
||||
|
||||
### GIFs and Images
|
||||
- New Claude Obsidian branded assets added (16x9 cover GIF, 1x1 GIF, static PNGs)
|
||||
- Compressed: `gif-cover-16x9.gif` 2.6MB→1.3MB (50%), `gif-1x1.gif` 2.6MB→848KB (68%) — via FFmpeg palette optimization, scaled to 960px/640px, 15fps, 128-color palette
|
||||
- Example screenshots added: `image-example-graph-view.png`, `image-example-wiki-map-view.png`
|
||||
|
||||
### README Structure (top to bottom)
|
||||
1. `claude-obsidian-gif-cover-16x9.gif` — header
|
||||
2. Description text
|
||||
3. `welcome-canvas.gif` — What It Does demo (full width)
|
||||
4. Descriptive paragraphs
|
||||
5. `image-example-graph-view.png` + `image-example-wiki-map-view.png` — side by side screenshots
|
||||
6. Quick Start → Commands → Cross-Project → Six Modes → What Gets Created → MCP → Plugins → CSS Snippets → Banner → File Structure → AutoResearch → Seed Vault
|
||||
7. `wiki-graph-grow.gif` + `workflow-loop.gif` — bottom of Seed Vault section
|
||||
|
||||
---
|
||||
|
||||
## Repository State
|
||||
|
||||
| Repo | Visibility | URL |
|
||||
|------|-----------|-----|
|
||||
| AI-Marketing-Hub/claude-obsidian | Public | https://github.com/AI-Marketing-Hub/claude-obsidian |
|
||||
| avalonreset-pro/claude-obsidian | Private | https://github.com/avalonreset-pro/claude-obsidian |
|
||||
|
||||
Community install command: `claude plugin install github:AI-Marketing-Hub/claude-obsidian`
|
||||
|
||||
Future updates: `git push origin main && git push community main`
|
||||
|
||||
---
|
||||
|
||||
## Key Decisions
|
||||
|
||||
- **Rename to claude-obsidian**: clearer branding, immediately communicates the Claude + Obsidian pairing
|
||||
- **avalonreset-pro repo is private**: community members only, not public
|
||||
- **Excalidraw main.js excluded from git**: 8MB downloaded by `setup-vault.sh` at setup time
|
||||
- **Bundled plugin redistribution**: acceptable — all 4 plugins are publicly distributed through Obsidian's community plugin system
|
||||
- **GIF compression strategy**: palette reduction (256→128 colors) + resolution scaling gives 50-68% savings with no visible quality loss at GitHub's rendering width
|
||||
@@ -0,0 +1,278 @@
|
||||
---
|
||||
type: session
|
||||
title: "claude-obsidian v1.4 Release Session"
|
||||
created: 2026-04-08
|
||||
updated: 2026-04-08
|
||||
tags:
|
||||
- meta
|
||||
- session
|
||||
- release
|
||||
- audit-response
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[claude-obsidian-ecosystem]]"
|
||||
- "[[cherry-picks]]"
|
||||
- "[[full-audit-and-system-setup-session]]"
|
||||
- "[[claude-obsidian-v1.2.0-release-session]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
sources:
|
||||
- "[[claude-obsidian-ecosystem-research]]"
|
||||
---
|
||||
|
||||
# claude-obsidian v1.4 Release Session
|
||||
|
||||
A complete release cycle covering v1.1, v1.4.0, and v1.4.1. Includes ecosystem research, external audit response, multi-agent compatibility rollout, full em-dash style cleanup, git history scrub for privacy, and a hotfix for the plugin install command syntax.
|
||||
|
||||
## Release Sequence
|
||||
|
||||
| Version | What shipped |
|
||||
|---|---|
|
||||
| v1.1 | URL ingestion, image/vision ingestion, delta tracking manifest, 3 new skills (defuddle, obsidian-bases, obsidian-markdown), multi-depth wiki-query modes, PostToolUse auto-commit hook, removal of invalid `allowed-tools` frontmatter field |
|
||||
| v1.4.0 | Dataview to Bases migration, Canvas JSON 1.0 spec completeness, hooks hardening plus PostCompact, MCP setup hardened with Obsidian CLI option, custom callouts documented, 6 multi-agent bootstrap files, 249 em dashes scrubbed, security-rewrote git history to remove a placeholder email |
|
||||
| v1.4.1 | Hotfix for the wrong plugin install command syntax in README and docs/install-guide.md |
|
||||
|
||||
## v1.1: First Feature Release of This Session
|
||||
|
||||
Shipped in response to an internal quality check against the wider ecosystem (16+ Claude plus Obsidian projects researched, filed in [[claude-obsidian-ecosystem]]). The highest-value features from competing implementations were cherry-picked and shipped as v1.1.
|
||||
|
||||
### New skills (Agent Skills spec compliant)
|
||||
|
||||
- `skills/defuddle/SKILL.md`: strips ads, nav, and clutter from web pages before URL ingestion. Saves 40 to 60 percent on tokens for typical articles.
|
||||
- `skills/obsidian-markdown/SKILL.md`: full reference for Obsidian Flavored Markdown (wikilinks, embeds, all callout types, properties, math, Mermaid). Cross-references kepano/obsidian-skills as the authoritative source.
|
||||
- `skills/obsidian-bases/SKILL.md`: native Obsidian Bases `.base` file syntax. Uses correct `filters/views/formulas` structure (NOT Dataview-style `from/where`). Rewritten mid-session after the first attempt used the wrong syntax.
|
||||
|
||||
### wiki-ingest upgrades
|
||||
|
||||
- **URL ingestion**: passes any `https://` URL directly. Uses WebFetch, optionally pipes through defuddle, saves to `.raw/articles/`, then runs the normal ingest pipeline.
|
||||
- **Image/vision ingestion**: `.png`, `.jpg`, `.gif`, `.webp`, etc. Claude reads the image natively, extracts text via OCR and concepts via vision, saves the description to `.raw/images/`, then ingests.
|
||||
- **Delta tracking**: `.raw/.manifest.json` tracks MD5 hash per source, timestamps, and the pages produced. Re-running ingest on unchanged files skips them automatically. Override with "force ingest".
|
||||
|
||||
### wiki-query multi-depth modes
|
||||
|
||||
Three query tiers:
|
||||
|
||||
- **Quick** (`query quick: ...`): hot.md plus index.md only. About 1,500 tokens. Best for fact lookups.
|
||||
- **Standard** (default): hot plus index plus 3 to 5 relevant pages. About 3,000 tokens. Best for most questions.
|
||||
- **Deep** (`query deep: ...`): full wiki cross-reference plus optional web search supplement. About 8,000+ tokens. Best for synthesis, comparisons, "tell me everything about X".
|
||||
|
||||
### Hook: PostToolUse auto-commit
|
||||
|
||||
Every `Write` or `Edit` tool call to `wiki/` or `.raw/` triggers `git add` and `git commit` automatically. Guarded by `[ -d .git ]` so it never errors in non-git directories, and by `git diff --cached --quiet` so it never creates empty commits. Matcher: `Write|Edit`.
|
||||
|
||||
### Critical fix: `allowed-tools` frontmatter removed
|
||||
|
||||
The Agent Skills spec only supports `name`, `description`, `argument-hint`, `compatibility`, `disable-model-invocation`, `license`, `metadata`, and `user-invokable` fields in SKILL.md frontmatter. The `allowed-tools` field was never valid and was being silently ignored. Removed from all 10 SKILL.md files to match the kepano/obsidian-skills authoritative convention.
|
||||
|
||||
## v1.4.0: External Audit Response
|
||||
|
||||
External auditor delivered a 21-source review (the "compass artifact") against Agent Skills spec, Claude Code hooks, Obsidian v1.9 through v1.12, and JSON Canvas 1.0. Initial audit score: 6.5/10. Many findings were already resolved in v1.1 (the audit was conducted against a snapshot from before that release). The remaining valid findings became v1.4.0. Full findings and prioritization are in [[cherry-picks]].
|
||||
|
||||
### Tier 1: Critical fixes
|
||||
|
||||
**Dataview to Bases migration** (the biggest correctness fix). Obsidian Bases shipped as a core plugin in v1.9.10 (August 2025), providing native database-like views that replace Dataview for most use cases. Created `wiki/meta/dashboard.base` with six views (Recent Activity, Seed Pages, Entities Missing Sources, Open Questions, Comparisons, Sources). Updated `wiki/meta/dashboard.md` to embed the new base file as primary and keep the legacy Dataview queries as optional fallback. Reorganized README plugins section to recommend Bases (core, no install needed) as primary and mark Dataview as optional/legacy.
|
||||
|
||||
**Canvas JSON 1.0 spec completeness**. Added previously missing fields to `skills/canvas/references/canvas-spec.md`:
|
||||
- Group nodes: `background` (string path) and `backgroundStyle` (`cover`, `ratio`, `repeat`)
|
||||
- Edges: `fromEnd` (defaults to `"none"`) and `toEnd` (defaults to `"arrow"`). Asymmetric defaults that produce a single arrow without explicit specification.
|
||||
- Documented the official hex ID convention alongside the descriptive ID alternative.
|
||||
|
||||
### Tier 2: Important improvements
|
||||
|
||||
**Hooks hardening plus PostCompact**. `hooks/hooks.json` updated:
|
||||
|
||||
- SessionStart now uses both `command` and `prompt` types. The command runs `[ -f wiki/hot.md ] && cat wiki/hot.md || true` as the canonical safety check that works in non-vault sessions without erroring. Matcher: `startup|resume`.
|
||||
- **NEW PostCompact hook** re-injects `wiki/hot.md` after context compaction. Critical insight: hook-injected context does NOT survive compaction, only `CLAUDE.md` does. Without this hook the hot cache disappears mid-session after any compact event.
|
||||
- PostToolUse auto-commit now guarded by `[ -d .git ]` in addition to the existing safeguards.
|
||||
- New `hooks/README.md` documents all four hooks plus the known plugin-hooks STDOUT bug (`anthropics/claude-code#10875`) and workarounds.
|
||||
|
||||
**MCP setup hardened**. `skills/wiki/references/mcp-setup.md` now has a `> [!warning]` callout above the `NODE_TLS_REJECT_UNAUTHORIZED: "0"` line explaining that it disables TLS verification process-wide (acceptable for `127.0.0.1` only). Added **Option D: Obsidian CLI** (Obsidian v1.12+) as the recommended alternative that avoids the TLS workaround entirely by using the native CLI via the Bash tool.
|
||||
|
||||
**Custom callouts documented**. The vault defines four custom callout types in `.obsidian/snippets/vault-colors.css`:
|
||||
|
||||
| Callout | Color | Icon | Use for |
|
||||
|---|---|---|---|
|
||||
| `contradiction` | reddish-brown | `lucide-alert-triangle` | New source conflicts with existing claim |
|
||||
| `gap` | beige | `lucide-help-circle` | Topic has no source yet |
|
||||
| `key-insight` | bright blue | `lucide-lightbulb` | Important takeaway worth highlighting |
|
||||
| `stale` | gray | `lucide-clock` | Claim may be outdated |
|
||||
|
||||
Full documentation added to `skills/wiki/references/css-snippets.md` including built-in fallback equivalents for users who do not want the custom CSS. `skills/wiki-ingest/SKILL.md` got an explicit note that `[!contradiction]` depends on the CSS snippet.
|
||||
|
||||
### Tier 3: Multi-agent compatibility (low complexity, high reach)
|
||||
|
||||
Skills are already in the cross-platform Agent Skills format. The only thing missing was adapter files so other AI coding agents could discover them. Added:
|
||||
|
||||
| File | For |
|
||||
|---|---|
|
||||
| `AGENTS.md` | Codex CLI, OpenCode |
|
||||
| `GEMINI.md` | Gemini CLI, Antigravity |
|
||||
| `.cursor/rules/claude-obsidian.mdc` | Cursor (always-on rules) |
|
||||
| `.windsurf/rules/claude-obsidian.md` | Windsurf Cascade |
|
||||
| `.github/copilot-instructions.md` | GitHub Copilot |
|
||||
| `bin/setup-multi-agent.sh` | Idempotent symlink installer that wires up `skills/` into each agent's expected location |
|
||||
|
||||
This turns claude-obsidian into a multi-agent plugin at near-zero compatibility cost. Pattern borrowed from [[Ar9av-obsidian-wiki]] which was the reference implementation for multi-agent support.
|
||||
|
||||
## Style Cleanup: Em Dash Scrub
|
||||
|
||||
Per user preference saved to feedback memory: never use em dashes (U+2014) or `--` as punctuation. Use periods, commas, colons, or parentheses instead. Hyphens in compound words (auto-commit, multi-agent) are fine.
|
||||
|
||||
Wrote a context-aware Python scrubber at `/tmp/scrub_em_dashes.py` with rules:
|
||||
|
||||
- Heading lines (`^#`): em dash becomes `:`
|
||||
- List items (`^-`, `^|`): em dash becomes `:` (for label-description patterns)
|
||||
- Prose: em dash becomes `.` with next word capitalized
|
||||
|
||||
**Result**: 249 em dashes removed across 26 files. Scrubbed every SKILL.md, every doc, every hook file, every bootstrap file, and marketplace.json. Manual smoothing required for:
|
||||
|
||||
- `skills/obsidian-markdown/SKILL.md`: 4 code-block annotation tables where the scrubber produced broken fragments. Converted to proper markdown tables.
|
||||
- `skills/wiki-query/SKILL.md`: 4 "If X. Respond." fragments rewritten as "If X, respond."
|
||||
- `bin/setup-multi-agent.sh`: 1 leftover em dash at end-of-line that the scrubber missed (only matched space-em-space). Plus one awkward echo string fixed.
|
||||
|
||||
The user-facing feedback was clear: "make it proper and natural". The scrubbed prose reads cleaner with the fragmented sentences smoothed out.
|
||||
|
||||
## Security: Email Removal and Git History Rewrite
|
||||
|
||||
A placeholder email `[scrubbed-email]` (which the user confirmed does not exist as a real address) was in `marketplace.json` plus two docs. Removed from working tree first, then rewrote git history to scrub it from all commits.
|
||||
|
||||
**Tool**: `git filter-repo` (available at `~/.pyenv/versions/3.12.4/bin/git-filter-repo`).
|
||||
|
||||
**Two passes required**:
|
||||
|
||||
1. `git filter-repo --replace-text /tmp/email-replacements.txt --force` scrubs blob contents across all commits.
|
||||
2. `git filter-repo --replace-message /tmp/email-msg-replacements.txt --force` scrubs commit messages. The first pass caught 3 occurrences in file contents but missed 1 occurrence in a commit subject line. The second pass caught that.
|
||||
|
||||
**Replacement string**: `[scrubbed-email]==>***REMOVED***`
|
||||
|
||||
**Post-rewrite actions**:
|
||||
- Re-added the `origin` remote that filter-repo removes for safety
|
||||
- Moved `v1.4.0` tag forward to include the security commit (since v1.4.0 had not been consumed yet by any user)
|
||||
- Force-pushed main plus both tags (`v1.1` and `v1.4.0`)
|
||||
- Updated the v1.4.0 GitHub release notes to include a "Security Note" section
|
||||
|
||||
**Verification**: grep across all refs, all blobs, all commit messages returned zero matches for the scrubbed email string. GitHub release bodies checked for same: both v1.1 and v1.4.0 release pages clean.
|
||||
|
||||
**Caveat for other clones**: history rewrite means every commit hash changed. Any other machine or private `community` remote that has the repo still contains the old history. Those need `git fetch && git reset --hard origin/main` or a force push to clean up.
|
||||
|
||||
## v1.4.1: Plugin Install Command Hotfix
|
||||
|
||||
The v1.4.0 README and install guide showed this install command:
|
||||
|
||||
```bash
|
||||
claude plugin install github:AI-Marketing-Hub/claude-obsidian
|
||||
```
|
||||
|
||||
This form does not exist in Claude Code. Users trying it see:
|
||||
|
||||
```
|
||||
Failed to install plugin "github:AI-Marketing-Hub/claude-obsidian": Plugin "github:AI-Marketing-Hub/claude-obsidian" not found in any configured marketplace
|
||||
```
|
||||
|
||||
### The correct install flow (per `code.claude.com/docs/en/plugin-marketplaces`)
|
||||
|
||||
Plugin installation is a **two-step** process:
|
||||
|
||||
```bash
|
||||
# Step 1: add the marketplace catalog
|
||||
claude plugin marketplace add AI-Marketing-Hub/claude-obsidian
|
||||
|
||||
# Step 2: install the plugin from the catalog by name
|
||||
claude plugin install claude-obsidian@claude-obsidian-marketplace
|
||||
```
|
||||
|
||||
Where `claude-obsidian` is the plugin name (from `plugin.json`) and `claude-obsidian-marketplace` is the marketplace name (from `marketplace.json`). The `@` delimiter separates them.
|
||||
|
||||
### Why the confusion existed
|
||||
|
||||
There is no `claude plugin install github:owner/repo` shortcut. The marketplace abstraction is mandatory: Claude Code always fetches via a registered marketplace. A single-repo plugin like claude-obsidian is both the marketplace host and the plugin host, and the user must register the marketplace first before installing any plugin from it.
|
||||
|
||||
### Related CLI commands (useful to know)
|
||||
|
||||
| Command | What it does |
|
||||
|---|---|
|
||||
| `claude plugin marketplace list` | Show all registered marketplaces |
|
||||
| `claude plugin marketplace add owner/repo` | Register a new marketplace from a GitHub repo |
|
||||
| `claude plugin marketplace update <name>` | Refresh the marketplace catalog and re-clone |
|
||||
| `claude plugin marketplace remove <name>` | Unregister a marketplace (also uninstalls its plugins) |
|
||||
| `claude plugin install <plugin>@<marketplace>` | Install a specific plugin |
|
||||
| `claude plugin list` | Show all installed plugins and their status |
|
||||
| `claude plugin validate .` | Validate a marketplace.json, plugin.json, and frontmatter |
|
||||
|
||||
### Files changed in v1.4.1
|
||||
|
||||
- `README.md`: Option 2 install section rewritten with two-step flow
|
||||
- `docs/install-guide.md`: same correction
|
||||
- `.claude-plugin/plugin.json`: 1.4.0 to 1.4.1
|
||||
- `.claude-plugin/marketplace.json`: both `metadata.version` and `plugins[0].version` bumped to 1.4.1
|
||||
|
||||
### Confirmed working
|
||||
|
||||
After v1.4.1 was published, the user ran the corrected commands and saw:
|
||||
|
||||
```
|
||||
claude-obsidian@claude-obsidian-marketplace
|
||||
Version: 1.4.1
|
||||
Scope: user
|
||||
Status: ✔ enabled
|
||||
```
|
||||
|
||||
v1.4.1 installed at user scope and enabled.
|
||||
|
||||
## Key Lessons (Worth Remembering)
|
||||
|
||||
1. **Plugin install is always two-step**. There is no github shorthand form. `marketplace add` then `install plugin@marketplace`.
|
||||
2. **`allowed-tools` is not a valid skill frontmatter field**. The Agent Skills spec only accepts `name`, `description`, `argument-hint`, `compatibility`, `disable-model-invocation`, `license`, `metadata`, `user-invokable`. kepano/obsidian-skills uses only `name` and `description` which is the gold standard convention.
|
||||
3. **Obsidian Bases uses `filters/views/formulas`, not Dataview-style `from/where`**. Easy to confuse. Always check `help.obsidian.md/bases/syntax` for the current syntax.
|
||||
4. **Canvas JSON 1.0 has asymmetric edge defaults**. `fromEnd` defaults to `"none"`, `toEnd` defaults to `"arrow"`. Omitting both produces a single arrow pointing from source to target.
|
||||
5. **Hook-injected context does not survive context compaction**. Only `CLAUDE.md` does. Any plugin that injects context via SessionStart hooks should also add a PostCompact hook to restore it mid-session.
|
||||
6. **`git filter-repo` needs two passes for full scrub**. `--replace-text` handles blob contents, `--replace-message` handles commit messages. Running only one leaves traces.
|
||||
7. **`git filter-repo` removes the `origin` remote for safety**. Must re-add it manually before force-pushing.
|
||||
8. **Marketplace name and plugin name can differ**. Our marketplace is `claude-obsidian-marketplace`, our plugin is `claude-obsidian`. The `@` delimiter disambiguates them.
|
||||
9. **Style preference: no em dashes anywhere**. Periods, commas, colons, or parentheses instead. Applies to all prose, commit messages, release notes, file content. Hyphens in compound words are fine.
|
||||
|
||||
## Files Created in This Session
|
||||
|
||||
Summary of everything new or newly created:
|
||||
|
||||
| Path | Type | Purpose |
|
||||
|---|---|---|
|
||||
| `skills/defuddle/SKILL.md` | skill | Web page cleaner |
|
||||
| `skills/obsidian-bases/SKILL.md` | skill | Obsidian Bases syntax |
|
||||
| `skills/obsidian-markdown/SKILL.md` | skill | Full Obsidian syntax reference |
|
||||
| `wiki/meta/dashboard.base` | bases dashboard | 6-view Bases dashboard |
|
||||
| `wiki/comparisons/claude-obsidian-ecosystem.md` | comparison | 16+ project feature matrix |
|
||||
| `wiki/concepts/cherry-picks.md` | concept | Prioritized feature backlog |
|
||||
| `wiki/sources/claude-obsidian-ecosystem-research.md` | source | Research summary |
|
||||
| `wiki/entities/Ar9av-obsidian-wiki.md` | entity | Multi-agent reference implementation |
|
||||
| `wiki/entities/Nexus-claudesidian-mcp.md` | entity | Native Obsidian plugin |
|
||||
| `wiki/entities/ballred-obsidian-claude-pkm.md` | entity | Goal cascade PKM |
|
||||
| `wiki/entities/rvk7895-llm-knowledge-bases.md` | entity | Multi-depth query reference |
|
||||
| `wiki/entities/kepano-obsidian-skills.md` | entity | Authoritative skill reference |
|
||||
| `wiki/entities/Claudian-YishenTu.md` | entity | Native Obsidian plugin |
|
||||
| `.raw/claude-obsidian-ecosystem-research.md` | raw source | Ecosystem research dump |
|
||||
| `hooks/README.md` | doc | Hook documentation |
|
||||
| `AGENTS.md` | bootstrap | Codex CLI / OpenCode |
|
||||
| `GEMINI.md` | bootstrap | Gemini CLI / Antigravity |
|
||||
| `.cursor/rules/claude-obsidian.mdc` | bootstrap | Cursor rules |
|
||||
| `.windsurf/rules/claude-obsidian.md` | bootstrap | Windsurf Cascade |
|
||||
| `.github/copilot-instructions.md` | bootstrap | GitHub Copilot |
|
||||
| `bin/setup-multi-agent.sh` | script | Multi-agent symlink installer |
|
||||
|
||||
## Current Plugin State
|
||||
|
||||
- **Plugin installed**: `claude-obsidian@claude-obsidian-marketplace` version `1.4.1`, user scope, enabled
|
||||
- **Releases on GitHub**: `v1.1`, `v1.4.0`, `v1.4.1`
|
||||
- **10 skills** in `skills/`: wiki, wiki-ingest, wiki-query, wiki-lint, save, autoresearch, canvas, defuddle, obsidian-bases, obsidian-markdown
|
||||
- **4 lifecycle hooks** in `hooks/hooks.json`: SessionStart, PostCompact, PostToolUse, Stop
|
||||
- **6 multi-agent bootstrap files** covering Codex, OpenCode, Gemini, Cursor, Windsurf, GitHub Copilot
|
||||
- **2 agents** in `agents/`: wiki-ingest, wiki-lint
|
||||
|
||||
## Deferred to v1.5.0
|
||||
|
||||
From the audit cherry-picks list, these items were identified but intentionally not included in v1.4.0:
|
||||
|
||||
- `/adopt` command for importing existing vaults (medium complexity, adds new surface)
|
||||
- Vault graph analysis enhancement to wiki-lint (hub pages, cross-domain bridges, dead-ends)
|
||||
- Semantic search via qmd MCP server (optional external dependency)
|
||||
- Marp slide output from wiki queries (niche)
|
||||
- Thinking mode vs Writing mode UX experiment
|
||||
@@ -0,0 +1,78 @@
|
||||
filters:
|
||||
and:
|
||||
- file.inFolder("wiki")
|
||||
- file.ext == "md"
|
||||
formulas:
|
||||
age_days: (now() - file.mtime).days
|
||||
properties:
|
||||
type:
|
||||
displayName: Type
|
||||
status:
|
||||
displayName: Status
|
||||
updated:
|
||||
displayName: Updated
|
||||
formula.age_days:
|
||||
displayName: Days Since Edit
|
||||
views:
|
||||
- type: table
|
||||
name: Recent Activity
|
||||
groupBy:
|
||||
property: type
|
||||
direction: ASC
|
||||
order:
|
||||
- file.name
|
||||
- type
|
||||
- status
|
||||
- updated
|
||||
- formula.age_days
|
||||
limit: 15
|
||||
- type: list
|
||||
name: Seed Pages (Need Development)
|
||||
filters:
|
||||
and:
|
||||
- status == "seed"
|
||||
order:
|
||||
- file.name
|
||||
- updated
|
||||
- type: list
|
||||
name: Entities Missing Sources
|
||||
filters:
|
||||
and:
|
||||
- file.inFolder("wiki/entities")
|
||||
- or:
|
||||
- "!sources"
|
||||
- sources.length == 0
|
||||
order:
|
||||
- file.name
|
||||
- type: list
|
||||
name: Open Questions
|
||||
filters:
|
||||
and:
|
||||
- file.inFolder("wiki/questions")
|
||||
- or:
|
||||
- status == "developing"
|
||||
- status == "seed"
|
||||
order:
|
||||
- file.name
|
||||
- updated
|
||||
- type: table
|
||||
name: Comparisons
|
||||
filters:
|
||||
and:
|
||||
- file.inFolder("wiki/comparisons")
|
||||
order:
|
||||
- file.name
|
||||
- verdict
|
||||
- updated
|
||||
- type: table
|
||||
name: Sources
|
||||
filters:
|
||||
and:
|
||||
- file.inFolder("wiki/sources")
|
||||
- type == "source"
|
||||
order:
|
||||
- file.name
|
||||
- author
|
||||
- date_published
|
||||
- updated
|
||||
limit: 10
|
||||
@@ -0,0 +1,68 @@
|
||||
---
|
||||
type: meta
|
||||
title: "Dashboard"
|
||||
updated: 2026-04-08
|
||||
tags:
|
||||
- meta
|
||||
- dashboard
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[index]]"
|
||||
- "[[overview]]"
|
||||
- "[[log]]"
|
||||
- "[[concepts/_index]]"
|
||||
- "[[Compounding Knowledge]]"
|
||||
---
|
||||
|
||||
# Wiki Dashboard
|
||||
|
||||
Navigation: [[index]] | [[overview]] | [[log]] | [[hot]]
|
||||
|
||||
The dashboard uses **Obsidian Bases**. A core Obsidian feature shipped in v1.9.10 (August 2025). No plugin install required.
|
||||
|
||||
> [!tip] Embedded Bases view
|
||||
> The interactive dashboard lives in [[dashboard.base]]. Open that file directly, or use the embed below.
|
||||
|
||||
![[dashboard.base]]
|
||||
|
||||
---
|
||||
|
||||
## Legacy Dataview Dashboard (Optional)
|
||||
|
||||
If you are on Obsidian < 1.9.10 or prefer Dataview, the queries below still work. Just install the Dataview community plugin.
|
||||
|
||||
### Recent Activity
|
||||
|
||||
```dataview
|
||||
TABLE type, status, updated FROM "wiki" SORT updated DESC LIMIT 15
|
||||
```
|
||||
|
||||
### Seed Pages (Need Development)
|
||||
|
||||
```dataview
|
||||
LIST FROM "wiki" WHERE status = "seed" SORT updated ASC
|
||||
```
|
||||
|
||||
### Entities Missing Sources
|
||||
|
||||
```dataview
|
||||
LIST FROM "wiki/entities" WHERE !sources OR length(sources) = 0
|
||||
```
|
||||
|
||||
### Open Questions
|
||||
|
||||
```dataview
|
||||
LIST FROM "wiki/questions" WHERE status = "developing" OR status = "seed" SORT updated DESC
|
||||
```
|
||||
|
||||
### Comparisons
|
||||
|
||||
```dataview
|
||||
TABLE verdict FROM "wiki/comparisons" SORT updated DESC
|
||||
```
|
||||
|
||||
### Sources
|
||||
|
||||
```dataview
|
||||
TABLE author, date_published, updated FROM "wiki/sources" WHERE type = "source" SORT updated DESC LIMIT 10
|
||||
```
|
||||
Binary file not shown.
|
After Width: | Height: | Size: 101 KiB |
@@ -0,0 +1,82 @@
|
||||
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 900 300" width="900" height="300" font-family="Space Grotesk, system-ui, -apple-system, sans-serif">
|
||||
<rect width="900" height="300" fill="#0A0A0A"/>
|
||||
<text x="450" y="36" fill="#F5F5F0" font-size="17" font-weight="700" text-anchor="middle">v1.6.0 end-to-end validation: six tests, all green</text>
|
||||
<text x="450" y="56" fill="#888888" font-size="11" text-anchor="middle">Teams orchestration, codex gpt-5.4 for sub-agents, chair for integration</text>
|
||||
|
||||
<!-- Horizontal flow, 6 nodes evenly spaced -->
|
||||
<!-- Centers at x = 90, 225, 360, 495, 630, 765; y = 150 -->
|
||||
<!-- Connecting lines -->
|
||||
<g stroke="#2D2D2D" stroke-width="1.5" fill="none">
|
||||
<line x1="132" y1="150" x2="183" y2="150"/>
|
||||
<line x1="267" y1="150" x2="318" y2="150"/>
|
||||
<line x1="402" y1="150" x2="453" y2="150"/>
|
||||
<line x1="537" y1="150" x2="588" y2="150"/>
|
||||
<line x1="672" y1="150" x2="723" y2="150"/>
|
||||
</g>
|
||||
|
||||
<!-- T0 -->
|
||||
<g>
|
||||
<circle cx="90" cy="150" r="42" fill="#111111" stroke="#16a34a" stroke-width="2"/>
|
||||
<text x="90" y="140" fill="#16a34a" font-size="10" font-weight="700" text-anchor="middle">T0</text>
|
||||
<text x="90" y="158" fill="#F5F5F0" font-size="14" font-weight="700" text-anchor="middle">ollama</text>
|
||||
<text x="90" y="172" fill="#888888" font-size="9" text-anchor="middle">pull model</text>
|
||||
<text x="90" y="218" fill="#888888" font-size="10" text-anchor="middle">nomic-embed-text</text>
|
||||
<text x="90" y="232" fill="#6a6a6a" font-size="9" text-anchor="middle">274 MB, 15 s</text>
|
||||
</g>
|
||||
|
||||
<!-- T1 -->
|
||||
<g>
|
||||
<circle cx="225" cy="150" r="42" fill="#111111" stroke="#16a34a" stroke-width="2"/>
|
||||
<text x="225" y="140" fill="#16a34a" font-size="10" font-weight="700" text-anchor="middle">T1</text>
|
||||
<text x="225" y="158" fill="#F5F5F0" font-size="14" font-weight="700" text-anchor="middle">M1</text>
|
||||
<text x="225" y="172" fill="#888888" font-size="9" text-anchor="middle">dry-run k=3</text>
|
||||
<text x="225" y="218" fill="#888888" font-size="10" text-anchor="middle">DRY-RUN OK</text>
|
||||
<text x="225" y="232" fill="#6a6a6a" font-size="9" text-anchor="middle">8 children, no em-dashes</text>
|
||||
</g>
|
||||
|
||||
<!-- T2 -->
|
||||
<g>
|
||||
<circle cx="360" cy="150" r="42" fill="#111111" stroke="#16a34a" stroke-width="2"/>
|
||||
<text x="360" y="140" fill="#16a34a" font-size="10" font-weight="700" text-anchor="middle">T2</text>
|
||||
<text x="360" y="158" fill="#F5F5F0" font-size="14" font-weight="700" text-anchor="middle">M2</text>
|
||||
<text x="360" y="172" fill="#888888" font-size="9" text-anchor="middle">allocate</text>
|
||||
<text x="360" y="218" fill="#888888" font-size="10" text-anchor="middle">counter 2 to 3</text>
|
||||
<text x="360" y="232" fill="#6a6a6a" font-size="9" text-anchor="middle">c-000002 reserved</text>
|
||||
</g>
|
||||
|
||||
<!-- T3 -->
|
||||
<g>
|
||||
<circle cx="495" cy="150" r="42" fill="#111111" stroke="#16a34a" stroke-width="2"/>
|
||||
<text x="495" y="140" fill="#16a34a" font-size="10" font-weight="700" text-anchor="middle">T3</text>
|
||||
<text x="495" y="158" fill="#F5F5F0" font-size="14" font-weight="700" text-anchor="middle">M3</text>
|
||||
<text x="495" y="172" fill="#888888" font-size="9" text-anchor="middle">full tiling</text>
|
||||
<text x="495" y="218" fill="#888888" font-size="10" text-anchor="middle">41 scanned, 21 embedded</text>
|
||||
<text x="495" y="232" fill="#6a6a6a" font-size="9" text-anchor="middle">0 errors, 15 review pairs</text>
|
||||
</g>
|
||||
|
||||
<!-- T4 -->
|
||||
<g>
|
||||
<circle cx="630" cy="150" r="42" fill="#111111" stroke="#16a34a" stroke-width="2"/>
|
||||
<text x="630" y="140" fill="#16a34a" font-size="10" font-weight="700" text-anchor="middle">T4</text>
|
||||
<text x="630" y="158" fill="#F5F5F0" font-size="14" font-weight="700" text-anchor="middle">M1</text>
|
||||
<text x="630" y="172" fill="#888888" font-size="9" text-anchor="middle">commit</text>
|
||||
<text x="630" y="218" fill="#888888" font-size="10" text-anchor="middle">fold committed</text>
|
||||
<text x="630" y="232" fill="#6a6a6a" font-size="9" text-anchor="middle">115 lines, 8 children</text>
|
||||
</g>
|
||||
|
||||
<!-- T6 -->
|
||||
<g>
|
||||
<circle cx="765" cy="150" r="42" fill="#111111" stroke="#16a34a" stroke-width="2"/>
|
||||
<text x="765" y="140" fill="#16a34a" font-size="10" font-weight="700" text-anchor="middle">T6</text>
|
||||
<text x="765" y="158" fill="#F5F5F0" font-size="14" font-weight="700" text-anchor="middle">M4</text>
|
||||
<text x="765" y="172" fill="#888888" font-size="9" text-anchor="middle">autoresearch</text>
|
||||
<text x="765" y="218" fill="#888888" font-size="10" text-anchor="middle">3 pages filed</text>
|
||||
<text x="765" y="232" fill="#6a6a6a" font-size="9" text-anchor="middle">6 URLs fetched, sourced</text>
|
||||
</g>
|
||||
|
||||
<!-- Verdict strip -->
|
||||
<g>
|
||||
<rect x="300" y="262" width="300" height="24" rx="12" fill="#16a34a" fill-opacity="0.15" stroke="#16a34a" stroke-width="1"/>
|
||||
<text x="450" y="279" fill="#4ADE80" font-size="11" font-weight="700" text-anchor="middle">ACCEPT · make test green (74+ assertions)</text>
|
||||
</g>
|
||||
</svg>
|
||||
|
After Width: | Height: | Size: 4.9 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 171 KiB |
@@ -0,0 +1,88 @@
|
||||
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 900 500" width="900" height="500" font-family="Space Grotesk, system-ui, -apple-system, sans-serif">
|
||||
<rect width="900" height="500" fill="#0A0A0A"/>
|
||||
<text x="450" y="36" fill="#F5F5F0" font-size="17" font-weight="700" text-anchor="middle">Boundary-first autoresearch in action</text>
|
||||
<text x="450" y="56" fill="#888888" font-size="11" text-anchor="middle">top-5 frontier from the 2026-04-24 run. Candidate picked, three pages filed.</text>
|
||||
|
||||
<!-- Left column: candidate list -->
|
||||
<g>
|
||||
<text x="60" y="100" fill="#E07850" font-size="11" font-weight="700" letter-spacing="2">TOP-5 FRONTIER</text>
|
||||
|
||||
<!-- #1 skipped -->
|
||||
<rect x="50" y="116" width="310" height="44" rx="10" fill="#111111" stroke="#2D2D2D" stroke-width="1"/>
|
||||
<text x="70" y="138" fill="#6a6a6a" font-size="10" text-anchor="start">1. Ecosystem Research</text>
|
||||
<text x="340" y="138" fill="#6a6a6a" font-size="10" text-anchor="end">4.693</text>
|
||||
<text x="70" y="152" fill="#dc2626" font-size="9">skipped: source page</text>
|
||||
|
||||
<!-- #2 skipped -->
|
||||
<rect x="50" y="168" width="310" height="44" rx="10" fill="#111111" stroke="#2D2D2D" stroke-width="1"/>
|
||||
<text x="70" y="190" fill="#6a6a6a" font-size="10">2. DragonScale Memory</text>
|
||||
<text x="340" y="190" fill="#6a6a6a" font-size="10" text-anchor="end">4.000</text>
|
||||
<text x="70" y="204" fill="#dc2626" font-size="9">skipped: self-reference</text>
|
||||
|
||||
<!-- #3 CHOSEN -->
|
||||
<rect x="50" y="220" width="310" height="58" rx="10" fill="#1A1A1A" stroke="#4ADE80" stroke-width="2"/>
|
||||
<text x="70" y="242" fill="#F5F5F0" font-size="12" font-weight="700">3. How does the LLM Wiki pattern work?</text>
|
||||
<text x="340" y="242" fill="#F5F5F0" font-size="11" font-weight="700" text-anchor="end">1.702</text>
|
||||
<text x="70" y="258" fill="#4ADE80" font-size="9" font-weight="700" letter-spacing="1">CHOSEN CANDIDATE</text>
|
||||
<text x="70" y="271" fill="#888888" font-size="9">question page with gap, 6 URLs fetched for extension</text>
|
||||
|
||||
<!-- #4 -->
|
||||
<rect x="50" y="286" width="310" height="44" rx="10" fill="#111111" stroke="#2D2D2D" stroke-width="1"/>
|
||||
<text x="70" y="308" fill="#888888" font-size="10">4. Wiki vs RAG</text>
|
||||
<text x="340" y="308" fill="#888888" font-size="10" text-anchor="end">1.135</text>
|
||||
<text x="70" y="322" fill="#6a6a6a" font-size="9">eligible, not picked this round</text>
|
||||
|
||||
<!-- #5 -->
|
||||
<rect x="50" y="338" width="310" height="44" rx="10" fill="#111111" stroke="#2D2D2D" stroke-width="1"/>
|
||||
<text x="70" y="360" fill="#888888" font-size="10">5. SEO Drift Monitoring</text>
|
||||
<text x="340" y="360" fill="#888888" font-size="10" text-anchor="end">0.717</text>
|
||||
<text x="70" y="374" fill="#6a6a6a" font-size="9">eligible, not picked this round</text>
|
||||
|
||||
<!-- Formula footer -->
|
||||
<text x="50" y="420" fill="#6a6a6a" font-size="10" font-weight="700">formula</text>
|
||||
<text x="50" y="436" fill="#888888" font-size="10">score = (out_degree - in_degree) * exp(-age_days / 30)</text>
|
||||
<text x="50" y="452" fill="#6a6a6a" font-size="9">halflife 30 days, no floor. 22 scoreable pages. meta excluded.</text>
|
||||
</g>
|
||||
|
||||
<!-- Arrow from chosen candidate into synthesis -->
|
||||
<g stroke="#4ADE80" stroke-width="2" fill="none">
|
||||
<path d="M 365 249 Q 440 249 520 200" marker-end="url(#arrow-green)"/>
|
||||
<path d="M 365 249 Q 440 249 520 249"/>
|
||||
<path d="M 365 249 Q 440 249 520 298" marker-end="url(#arrow-green)"/>
|
||||
<line x1="519" y1="249" x2="518" y2="250"/>
|
||||
</g>
|
||||
<defs>
|
||||
<marker id="arrow-green" viewBox="0 0 10 10" refX="9" refY="5" markerWidth="6" markerHeight="6" orient="auto">
|
||||
<path d="M 0 0 L 10 5 L 0 10 z" fill="#4ADE80"/>
|
||||
</marker>
|
||||
</defs>
|
||||
|
||||
<!-- Right column: 3 new pages filed -->
|
||||
<g>
|
||||
<text x="540" y="100" fill="#4ADE80" font-size="11" font-weight="700" letter-spacing="2">PAGES FILED</text>
|
||||
|
||||
<!-- New page 1 -->
|
||||
<rect x="520" y="176" width="340" height="54" rx="10" fill="#111111" stroke="#4ADE80" stroke-width="1.5"/>
|
||||
<text x="540" y="198" fill="#F5F5F0" font-size="12" font-weight="700">Persistent Wiki Artifact</text>
|
||||
<text x="540" y="214" fill="#888888" font-size="9">44 lines · Karpathy gist + Obsidian docs</text>
|
||||
<text x="540" y="226" fill="#6a6a6a" font-size="9">durable Markdown layer as the agent's memory object</text>
|
||||
|
||||
<!-- New page 2 -->
|
||||
<rect x="520" y="238" width="340" height="54" rx="10" fill="#111111" stroke="#4ADE80" stroke-width="1.5"/>
|
||||
<text x="540" y="260" fill="#F5F5F0" font-size="12" font-weight="700">Source-First Synthesis</text>
|
||||
<text x="540" y="276" fill="#888888" font-size="9">42 lines · Karpathy gist + RAG arXiv 2005.11401</text>
|
||||
<text x="540" y="288" fill="#6a6a6a" font-size="9">raw sources immutable, synthesis cites them</text>
|
||||
|
||||
<!-- New page 3 -->
|
||||
<rect x="520" y="300" width="340" height="54" rx="10" fill="#111111" stroke="#4ADE80" stroke-width="1.5"/>
|
||||
<text x="540" y="322" fill="#F5F5F0" font-size="12" font-weight="700">Query-Time Retrieval</text>
|
||||
<text x="540" y="338" fill="#888888" font-size="9">44 lines · Karpathy gist + MemGPT arXiv 2310.08560</text>
|
||||
<text x="540" y="350" fill="#6a6a6a" font-size="9">wiki query synthesis vs Obsidian in-vault search</text>
|
||||
|
||||
<!-- Agenda-control caveat -->
|
||||
<rect x="520" y="400" width="340" height="60" rx="10" fill="#1A1A1A" stroke="#E07850" stroke-width="1"/>
|
||||
<text x="540" y="420" fill="#E07850" font-size="10" font-weight="700" letter-spacing="2">AGENDA CONTROL</text>
|
||||
<text x="540" y="436" fill="#888888" font-size="9">M4 does not just describe the vault; it shapes what</text>
|
||||
<text x="540" y="450" fill="#888888" font-size="9">the agent researches next. User picks, overrides, declines.</text>
|
||||
</g>
|
||||
</svg>
|
||||
|
After Width: | Height: | Size: 5.7 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 136 KiB |
@@ -0,0 +1,56 @@
|
||||
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 800 500" width="800" height="500" font-family="Space Grotesk, system-ui, -apple-system, sans-serif">
|
||||
<rect width="800" height="500" fill="#0A0A0A"/>
|
||||
<text x="400" y="36" fill="#F5F5F0" font-size="17" font-weight="700" text-anchor="middle">DragonScale Memory: four mechanisms</text>
|
||||
<text x="400" y="56" fill="#888888" font-size="11" font-weight="400" text-anchor="middle">opt-in, feature-gated, v1.6.0</text>
|
||||
|
||||
<!-- Card 1: Fold operator (stand-alone, no shared gate) -->
|
||||
<g>
|
||||
<rect x="40" y="90" width="355" height="140" rx="16" fill="#111111" stroke="#2D2D2D" stroke-width="1.5"/>
|
||||
<rect x="40" y="90" width="355" height="4" rx="2" fill="#E07850"/>
|
||||
<text x="60" y="120" fill="#E07850" font-size="13" font-weight="700" letter-spacing="2">M1</text>
|
||||
<text x="60" y="146" fill="#F5F5F0" font-size="15" font-weight="700">Fold operator</text>
|
||||
<text x="60" y="168" fill="#888888" font-size="10">Flat extractive rollup of the last 2^k log entries</text>
|
||||
<text x="60" y="184" fill="#888888" font-size="10">into a fold page. Dry-run default, commit explicit.</text>
|
||||
<text x="60" y="212" fill="#6a6a6a" font-size="9">skills/wiki-fold · dry-run stdout-only, no residue</text>
|
||||
</g>
|
||||
|
||||
<!-- Card 2: Addresses -->
|
||||
<g>
|
||||
<rect x="405" y="90" width="355" height="140" rx="16" fill="#111111" stroke="#2D2D2D" stroke-width="1.5"/>
|
||||
<rect x="405" y="90" width="355" height="4" rx="2" fill="#60A5FA"/>
|
||||
<text x="425" y="120" fill="#60A5FA" font-size="13" font-weight="700" letter-spacing="2">M2</text>
|
||||
<text x="425" y="146" fill="#F5F5F0" font-size="15" font-weight="700">Deterministic page addresses</text>
|
||||
<text x="425" y="168" fill="#888888" font-size="10">Creation-order counter c-NNNNNN, flock-guarded,</text>
|
||||
<text x="425" y="184" fill="#888888" font-size="10">recovers from max observed. Gap-tolerant.</text>
|
||||
<text x="425" y="212" fill="#6a6a6a" font-size="9">scripts/allocate-address.sh · counter at 3 in this vault</text>
|
||||
</g>
|
||||
|
||||
<!-- Card 3: Tiling -->
|
||||
<g>
|
||||
<rect x="40" y="250" width="355" height="140" rx="16" fill="#111111" stroke="#2D2D2D" stroke-width="1.5"/>
|
||||
<rect x="40" y="250" width="355" height="4" rx="2" fill="#8b5cf6"/>
|
||||
<text x="60" y="280" fill="#8b5cf6" font-size="13" font-weight="700" letter-spacing="2">M3</text>
|
||||
<text x="60" y="306" fill="#F5F5F0" font-size="15" font-weight="700">Semantic tiling lint</text>
|
||||
<text x="60" y="328" fill="#888888" font-size="10">Embedding-based duplicate detection via local</text>
|
||||
<text x="60" y="344" fill="#888888" font-size="10">ollama + nomic-embed-text. Banded thresholds.</text>
|
||||
<text x="60" y="372" fill="#6a6a6a" font-size="9">scripts/tiling-check.py · first run: 0 errors, 15 review pairs</text>
|
||||
</g>
|
||||
|
||||
<!-- Card 4: Boundary -->
|
||||
<g>
|
||||
<rect x="405" y="250" width="355" height="140" rx="16" fill="#111111" stroke="#2D2D2D" stroke-width="1.5"/>
|
||||
<rect x="405" y="250" width="355" height="4" rx="2" fill="#4ADE80"/>
|
||||
<text x="425" y="280" fill="#4ADE80" font-size="13" font-weight="700" letter-spacing="2">M4</text>
|
||||
<text x="425" y="306" fill="#F5F5F0" font-size="15" font-weight="700">Boundary-first autoresearch</text>
|
||||
<text x="425" y="328" fill="#888888" font-size="10">score = (out_degree - in_degree) * exp(-days/30).</text>
|
||||
<text x="425" y="344" fill="#888888" font-size="10">Agenda control, not pure memory. User picks.</text>
|
||||
<text x="425" y="372" fill="#6a6a6a" font-size="9">scripts/boundary-score.py · stdout-only, read-only</text>
|
||||
</g>
|
||||
|
||||
<!-- Shared gate strip for M2/M3/M4 -->
|
||||
<g>
|
||||
<rect x="40" y="420" width="720" height="44" rx="10" fill="#1A1A1A" stroke="#2D2D2D" stroke-width="1"/>
|
||||
<text x="60" y="440" fill="#E07850" font-size="11" font-weight="700" letter-spacing="2">SHARED GATE</text>
|
||||
<text x="60" y="458" fill="#888888" font-size="10">.vault-meta/ presence enables M2, M3, M4 feature detection. M1 has no shared state. Remove .vault-meta/ to disable all three at once.</text>
|
||||
</g>
|
||||
</svg>
|
||||
|
After Width: | Height: | Size: 4.0 KiB |
@@ -0,0 +1,91 @@
|
||||
---
|
||||
type: session
|
||||
title: "Full Audit, System Setup & Plugin Installation"
|
||||
created: 2026-04-07
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- session
|
||||
- audit
|
||||
- setup
|
||||
- plugin-install
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[claude-obsidian-v1.2.0-release-session]]"
|
||||
- "[[getting-started]]"
|
||||
- "[[index]]"
|
||||
---
|
||||
|
||||
# Full Audit, System Setup & Plugin Installation
|
||||
|
||||
Post-release audit session. Covers comprehensive repo audit across 12 areas, 3 issue fixes, plugin installation into the local Claude Code system, folder rename, and memory save.
|
||||
|
||||
---
|
||||
|
||||
## Audit Results (12 Areas)
|
||||
|
||||
All 12 areas audited — 3 issues found, all fixed same session.
|
||||
|
||||
### Issues Found and Fixed
|
||||
|
||||
| Issue | Fix |
|
||||
|-------|-----|
|
||||
| `Cosmic Brain Clean.gif` tracked in git (personal asset) | Removed with `git rm --cached`, added `Cosmic Brain*.gif` to .gitignore |
|
||||
| `Cosmic Brain Cover.png` tracked in git (personal asset) | Removed with `git rm --cached`, added `Cosmic Brain*.png` to .gitignore |
|
||||
| `Welcome.md` tracked in git (Obsidian personal file) | Removed with `git rm --cached`, added `Welcome.md` to .gitignore |
|
||||
| `vault-colors.css` comment said "cosmic-brain vault colors" | Updated to "claude-obsidian vault colors" |
|
||||
| `docs/superpowers/plans/` not committed | Committed audit plan file |
|
||||
|
||||
### Clean Areas (no issues)
|
||||
- Plugin manifests: all fields correct, version 1.2.0 consistent
|
||||
- All 7 SKILL.md files: valid frontmatter, correct tools, complete instructions
|
||||
- All 4 commands: mapped to correct skills, descriptions accurate
|
||||
- Both agents: model/maxTurns/tools correct
|
||||
- hooks/hooks.json: valid JSON, SessionStart + Stop hooks correct
|
||||
- All .obsidian/*.json: community-plugins.json (4 entries), appearance.json (3 snippets), app.json, graph.json all valid
|
||||
- All 4 Obsidian plugin manifests: complete, no personal data in data.json files
|
||||
- All 3 CSS snippets: GPL-2.0 headers present, no stale references
|
||||
- All 16 wikilinks resolve to valid files
|
||||
- All 3 canvases valid JSON, no broken file node references
|
||||
- README: all 6 images verified on disk, install commands correct, structure accurate
|
||||
- Zero secrets in any tracked file, all API key references are placeholders
|
||||
- Install simulation: all 7 skills, 4 commands, 2 agents discoverable, hooks valid
|
||||
|
||||
---
|
||||
|
||||
## Plugin Installation
|
||||
|
||||
claude-obsidian is now installed in the local Claude Code system:
|
||||
|
||||
```bash
|
||||
# Registered as marketplace
|
||||
claude plugin marketplace add AI-Marketing-Hub/claude-obsidian
|
||||
# → claude-obsidian-marketplace registered (user scope)
|
||||
|
||||
# Installed plugin
|
||||
claude plugin install claude-obsidian
|
||||
# → claude-obsidian@claude-obsidian-marketplace (scope: user) ✓
|
||||
```
|
||||
|
||||
To verify: `claude plugin list | grep claude-obsidian`
|
||||
|
||||
---
|
||||
|
||||
## System State
|
||||
|
||||
- Plugin repo: `~/claude-obsidian/` (git repo, both remotes live)
|
||||
- Plugin installed: `claude-obsidian@claude-obsidian-marketplace` (user scope, enabled)
|
||||
- Working folder renamed: `~/Desktop/Obsidian & Claude/` → `~/Desktop/claude-obsidian/`
|
||||
- Karpathy Gist comment drafted (ready to post at gist.github.com/karpathy/442a6bf555914893e9891c11519de94f)
|
||||
|
||||
---
|
||||
|
||||
## Commands Available After Install
|
||||
|
||||
| Trigger | What happens |
|
||||
|---------|-------------|
|
||||
| `/wiki` | Setup check, scaffold, or continue |
|
||||
| `ingest [file]` | Create 8–15 wiki pages from source |
|
||||
| `/save` | File this conversation to wiki |
|
||||
| `/autoresearch [topic]` | Autonomous web research loop |
|
||||
| `/canvas` | Visual canvas operations |
|
||||
| `lint the wiki` | Health check |
|
||||
Binary file not shown.
|
After Width: | Height: | Size: 351 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 212 KiB |
@@ -0,0 +1,402 @@
|
||||
---
|
||||
type: meta
|
||||
title: "v1.7 Retrieval Benchmark Corpus"
|
||||
status: evergreen
|
||||
updated: 2026-05-17
|
||||
---
|
||||
|
||||
# v1.7 Retrieval Benchmark — 50 Queries
|
||||
|
||||
Used by the v1.7.0 audit (`docs/audits/v1.7.0-audit-2026-05-17.md`) to score
|
||||
top-1 / top-5 accuracy of the hybrid retrieval pipeline (`scripts/retrieve.py`)
|
||||
vs. the simulated v1.6 baseline (`scripts/baseline-v16.py`).
|
||||
|
||||
## Schema
|
||||
|
||||
Each query: `id`, `query`, `correct` (canonical page paths), `relevant`
|
||||
(supporting paths), `category`, `rationale`. For negative queries, `correct`
|
||||
is `null` and success is defined as "returns nothing OR returns only pages
|
||||
in `relevant`."
|
||||
|
||||
Machine-parsable: every query block is bracketed by `### <id>` and ends
|
||||
before the next `### ` or `## `. The scoring runner at
|
||||
`scripts/benchmark-runner.py` reads this file directly.
|
||||
|
||||
## Scoring rules
|
||||
|
||||
- **top-1 success**: top result's `page_path` equals one of `correct[]`
|
||||
- **top-5 success**: any of top-5 results' `page_path` is in `correct[]`
|
||||
- For negative queries (correct=null): top-1 success if top result in `relevant[]` or no results; top-5 success if any top-5 in `relevant[]` or no results
|
||||
|
||||
Final scores reported as (top-1%, top-5%) per pipeline, broken down by category.
|
||||
|
||||
---
|
||||
|
||||
## Derived Queries (25)
|
||||
|
||||
### D1
|
||||
- query: How does knowledge compound differently in a wiki versus traditional RAG?
|
||||
- correct: wiki/concepts/Compounding Knowledge.md
|
||||
- relevant: wiki/concepts/LLM Wiki Pattern.md, wiki/comparisons/Wiki vs RAG.md
|
||||
- category: derived
|
||||
- rationale: Canonical treatment of why wiki accumulation produces more value over time vs. query-time re-derivation.
|
||||
|
||||
### D2
|
||||
- query: What is the Hot Cache and why would a user enable it in their session?
|
||||
- correct: wiki/concepts/Hot Cache.md
|
||||
- relevant: wiki/concepts/LLM Wiki Pattern.md, wiki/getting-started.md
|
||||
- category: derived
|
||||
- rationale: Direct explanation of the session context mechanism and its token-cost benefits.
|
||||
|
||||
### D3
|
||||
- query: Tell me the three-layer architecture behind the LLM Wiki pattern.
|
||||
- correct: wiki/concepts/LLM Wiki Pattern.md
|
||||
- relevant: wiki/overview.md, wiki/index.md
|
||||
- category: derived
|
||||
- rationale: Explicit description of .raw, wiki, and CLAUDE.md layers.
|
||||
|
||||
### D4
|
||||
- query: What does the wiki fold operator do and why use exponential batch sizes?
|
||||
- correct: wiki/concepts/DragonScale Memory.md
|
||||
- relevant: wiki/log.md, wiki/hot.md
|
||||
- category: derived
|
||||
- rationale: Mechanism 1 detailed explanation with 2^k justification and LSM analogy.
|
||||
|
||||
### D5
|
||||
- query: How are deterministic page addresses assigned and what format do they use?
|
||||
- correct: wiki/concepts/DragonScale Memory.md
|
||||
- relevant: wiki/log.md
|
||||
- category: derived
|
||||
- rationale: Mechanism 2 spec with c-NNNNNN format and counter recovery rules.
|
||||
|
||||
### D6
|
||||
- query: Explain semantic tiling lint and what cosine similarity threshold it uses.
|
||||
- correct: wiki/concepts/DragonScale Memory.md
|
||||
- relevant: wiki/log.md, wiki/hot.md
|
||||
- category: derived
|
||||
- rationale: Mechanism 3 detailed treatment including banded thresholds and embeddings.
|
||||
|
||||
### D7
|
||||
- query: What are the cherry-picked features claude-obsidian should implement next?
|
||||
- correct: wiki/concepts/cherry-picks.md
|
||||
- relevant: wiki/index.md, wiki/comparisons/claude-obsidian-ecosystem.md
|
||||
- category: derived
|
||||
- rationale: Explicit prioritized backlog from ecosystem research across tiers.
|
||||
|
||||
### D8
|
||||
- query: What is the relationship between SVG diagrams and brand tokens in claude-obsidian assets?
|
||||
- correct: wiki/concepts/SVG Diagram Style Guide.md
|
||||
- relevant: wiki/log.md
|
||||
- category: derived
|
||||
- rationale: Canonical style reference with color palette, typography, and layout tokens.
|
||||
|
||||
### D9
|
||||
- query: How does a user ingest a source document into claude-obsidian and what happens next?
|
||||
- correct: wiki/getting-started.md
|
||||
- relevant: wiki/concepts/LLM Wiki Pattern.md, wiki/overview.md
|
||||
- category: derived
|
||||
- rationale: Three-step quick start walkthrough with expected output pages.
|
||||
|
||||
### D10
|
||||
- query: What does the Pro Hub Challenge reward and what was the first winner?
|
||||
- correct: wiki/concepts/Pro Hub Challenge.md
|
||||
- relevant: wiki/entities/Claude SEO.md, wiki/log.md
|
||||
- category: derived
|
||||
- rationale: Community challenge structure and integration pattern with historical results.
|
||||
|
||||
### D11
|
||||
- query: What is Search Experience Optimization and how does it read SERPs backwards?
|
||||
- correct: wiki/concepts/Search Experience Optimization.md
|
||||
- relevant: wiki/entities/Claude SEO.md, wiki/concepts/Semantic Topic Clustering.md
|
||||
- category: derived
|
||||
- rationale: Direct explanation of SXO methodology and its page-type mismatch detection.
|
||||
|
||||
### D12
|
||||
- query: How does semantic topic clustering use overlapping URLs to detect content relationships?
|
||||
- correct: wiki/concepts/Semantic Topic Clustering.md
|
||||
- relevant: wiki/entities/Claude SEO.md, wiki/log.md
|
||||
- category: derived
|
||||
- rationale: Overlap scoring rules and hub-spoke architecture.
|
||||
|
||||
### D13
|
||||
- query: What tracking rules does SEO Drift Monitoring enforce and where is the baseline stored?
|
||||
- correct: wiki/concepts/SEO Drift Monitoring.md
|
||||
- relevant: wiki/entities/Claude SEO.md
|
||||
- category: derived
|
||||
- rationale: 17 comparison rules across severity levels and SQLite persistence location.
|
||||
|
||||
### D14
|
||||
- query: Who is Andrej Karpathy and why does he matter to this wiki project?
|
||||
- correct: wiki/entities/Andrej Karpathy.md
|
||||
- relevant: wiki/concepts/LLM Wiki Pattern.md, wiki/index.md
|
||||
- category: derived
|
||||
- rationale: Founder attribution and origination of the pattern with notable quote.
|
||||
|
||||
### D15
|
||||
- query: What version of Claude SEO shipped in April 2026 and how many skills does it have?
|
||||
- correct: wiki/entities/Claude SEO.md
|
||||
- relevant: wiki/log.md, wiki/meta/2026-04-14-claude-seo-v190-session.md
|
||||
- category: derived
|
||||
- rationale: v1.9.0 state with skill counts and Pro Hub Challenge integration.
|
||||
|
||||
### D16
|
||||
- query: Does Ar9av's obsidian-wiki support multi-agent deployment and how?
|
||||
- correct: wiki/entities/Ar9av-obsidian-wiki.md
|
||||
- relevant: wiki/comparisons/claude-obsidian-ecosystem.md, wiki/concepts/cherry-picks.md
|
||||
- category: derived
|
||||
- rationale: setup.sh bootstrap matrix across multiple agents and delta tracking manifest.
|
||||
|
||||
### D17
|
||||
- query: What does the /adopt command in ballred's PKM system do?
|
||||
- correct: wiki/entities/ballred-obsidian-claude-pkm.md
|
||||
- relevant: wiki/concepts/cherry-picks.md, wiki/comparisons/claude-obsidian-ecosystem.md
|
||||
- category: derived
|
||||
- rationale: Vault analysis and pattern detection for PARA, Zettelkasten, LYT migration.
|
||||
|
||||
### D18
|
||||
- query: What official Obsidian skills does kepano publish and which is most token-efficient?
|
||||
- correct: wiki/entities/kepano-obsidian-skills.md
|
||||
- relevant: wiki/concepts/cherry-picks.md, wiki/entities/Claudian-YishenTu.md
|
||||
- category: derived
|
||||
- rationale: obsidian-markdown, obsidian-bases, defuddle, and defuddle's 40-60% token reduction.
|
||||
|
||||
### D19
|
||||
- query: What are the three query depths in rvk7895's LLM knowledge base system?
|
||||
- correct: wiki/entities/rvk7895-llm-knowledge-bases.md
|
||||
- relevant: wiki/concepts/cherry-picks.md, wiki/comparisons/claude-obsidian-ecosystem.md
|
||||
- category: derived
|
||||
- rationale: Quick, Standard, Deep modes with output formats.
|
||||
|
||||
### D20
|
||||
- query: How does Nexus MCP for Obsidian store workspace memory and is it Obsidian Sync compatible?
|
||||
- correct: wiki/entities/Nexus-claudesidian-mcp.md
|
||||
- relevant: wiki/comparisons/claude-obsidian-ecosystem.md
|
||||
- category: derived
|
||||
- rationale: JSONL storage and Sync auto-inclusion.
|
||||
|
||||
### D21
|
||||
- query: What makes Claudian's inline edit feature best-in-class compared to other Obsidian AI plugins?
|
||||
- correct: wiki/entities/Claudian-YishenTu.md
|
||||
- relevant: wiki/comparisons/claude-obsidian-ecosystem.md
|
||||
- category: derived
|
||||
- rationale: Word-level diff preview and one-click apply with Plan Mode support.
|
||||
|
||||
### D22
|
||||
- query: Between claude-obsidian and other projects, who has the hot cache mechanism?
|
||||
- correct: wiki/comparisons/claude-obsidian-ecosystem.md
|
||||
- relevant: wiki/concepts/Hot Cache.md, wiki/index.md
|
||||
- category: derived
|
||||
- rationale: Feature matrix shows claude-obsidian is unique in session context caching.
|
||||
|
||||
### D23
|
||||
- query: When was the first real fold committed to the vault and what was its fold_id?
|
||||
- correct: wiki/folds/fold-k3-from-2026-04-23-to-2026-04-24-n8.md
|
||||
- relevant: wiki/log.md, wiki/hot.md, wiki/concepts/DragonScale Memory.md
|
||||
- category: derived
|
||||
- rationale: Fold metadata with exact creation date and 8-entry batch documentation.
|
||||
|
||||
### D24
|
||||
- query: What pages were created after the boundary-first autoresearch frontier pass?
|
||||
- correct: wiki/log.md
|
||||
- relevant: wiki/concepts/Persistent Wiki Artifact.md, wiki/concepts/Source-First Synthesis.md, wiki/concepts/Query-Time Retrieval.md
|
||||
- category: derived
|
||||
- rationale: Log entry for 2026-04-24 M4 run explicitly lists the three new concept pages filed.
|
||||
|
||||
### D25
|
||||
- query: What is a Persistent Wiki Artifact and how does it differ from ephemeral chat turns?
|
||||
- correct: wiki/concepts/Persistent Wiki Artifact.md
|
||||
- relevant: wiki/concepts/LLM Wiki Pattern.md, wiki/concepts/Source-First Synthesis.md
|
||||
- category: derived
|
||||
- rationale: Explicit definition of durable memory object with frontmatter, title, wikilinks, and provenance.
|
||||
|
||||
---
|
||||
|
||||
## Hard Queries (25)
|
||||
|
||||
### H1
|
||||
- query: How does the exponential compaction operator summarize batch entries?
|
||||
- correct: wiki/concepts/DragonScale Memory.md
|
||||
- relevant: wiki/log.md
|
||||
- category: synonym
|
||||
- rationale: "Fold operator" and "rollup" are interchangeable; query uses LSM terminology instead of dragon-curve naming.
|
||||
|
||||
### H2
|
||||
- query: What stable identifiers help page references survive renames in prompt-cache workloads?
|
||||
- correct: wiki/concepts/DragonScale Memory.md
|
||||
- relevant: wiki/log.md, wiki/hot.md
|
||||
- category: synonym
|
||||
- rationale: Deterministic addresses (c-NNNNNN format) are the answer; query uses prompt-cache context instead.
|
||||
|
||||
### H3
|
||||
- query: How does the semantic tiling mechanism detect and flag duplicate or near-duplicate pages?
|
||||
- correct: wiki/concepts/DragonScale Memory.md
|
||||
- relevant: wiki/log.md, wiki/hot.md
|
||||
- category: synonym
|
||||
- rationale: Mechanism 3 performs dedup via cosine similarity; query uses "near-duplicate" instead of "semantic tiling."
|
||||
|
||||
### H4
|
||||
- query: When user asks /autoresearch without a topic, which pages are candidates and how are they scored?
|
||||
- correct: wiki/concepts/DragonScale Memory.md
|
||||
- relevant: wiki/log.md, wiki/hot.md
|
||||
- category: synonym
|
||||
- rationale: Boundary-first autoresearch (Mechanism 4) surfaces frontier pages; query uses "autoresearch no-topic" instead.
|
||||
|
||||
### H5
|
||||
- query: How should the wiki layer stay connected to raw source material without replacing source content?
|
||||
- correct: wiki/concepts/Source-First Synthesis.md
|
||||
- relevant: wiki/concepts/LLM Wiki Pattern.md, wiki/concepts/Persistent Wiki Artifact.md
|
||||
- category: synonym
|
||||
- rationale: Source-first synthesis is the provenance rule; query asks for the principle without using the term.
|
||||
|
||||
### H6
|
||||
- query: When building a wiki versus RAG, how does the cost structure differ in maintenance, infrastructure, and long-term knowledge value?
|
||||
- correct: wiki/concepts/LLM Wiki Pattern.md, wiki/comparisons/Wiki vs RAG.md, wiki/concepts/Compounding Knowledge.md
|
||||
- relevant: wiki/overview.md
|
||||
- category: cross-page
|
||||
- rationale: Spans three concepts: architecture table, comparison table, and compounding mechanics.
|
||||
|
||||
### H7
|
||||
- query: What does the wiki-fold skill output in dry-run mode and how does it create child references?
|
||||
- correct: wiki/concepts/DragonScale Memory.md, wiki/log.md
|
||||
- relevant: wiki/hot.md
|
||||
- category: cross-page
|
||||
- rationale: Mechanism 1 spec + dry-run artifact from 2026-04-24 Phase 1 log entry cross-reference.
|
||||
|
||||
### H8
|
||||
- query: How would adding URL ingestion and web cleaning together reduce token usage during ingest?
|
||||
- correct: wiki/concepts/cherry-picks.md, wiki/entities/kepano-obsidian-skills.md
|
||||
- relevant: wiki/getting-started.md
|
||||
- category: cross-page
|
||||
- rationale: Combining cherry-pick #1 (URL ingest) and defuddle (40-60% token reduction) from two sources.
|
||||
|
||||
### H9
|
||||
- query: Which project has the best auto-commit hook and how did that feature get integrated into Claude SEO?
|
||||
- correct: wiki/entities/ballred-obsidian-claude-pkm.md, wiki/entities/Claude SEO.md, wiki/log.md
|
||||
- relevant: wiki/comparisons/claude-obsidian-ecosystem.md
|
||||
- category: cross-page
|
||||
- rationale: ballred has PostToolUse auto-commit; Claude SEO v1.9.0 release cites integration patterns.
|
||||
|
||||
### H10
|
||||
- query: How would you audit a page for SERP alignment, topic cannibalization risk, and performance regressions together?
|
||||
- correct: wiki/concepts/Search Experience Optimization.md, wiki/concepts/Semantic Topic Clustering.md, wiki/concepts/SEO Drift Monitoring.md
|
||||
- relevant: wiki/entities/Claude SEO.md, wiki/log.md
|
||||
- category: cross-page
|
||||
- rationale: Three separate v1.9.0 features; query asks for integrated workflow across all three.
|
||||
|
||||
### H11
|
||||
- query: What were the six test stages and did all mechanisms pass validation before v1.6.0 shipped?
|
||||
- correct: wiki/log.md, wiki/hot.md
|
||||
- relevant: wiki/concepts/DragonScale Memory.md
|
||||
- category: cross-page
|
||||
- rationale: 2026-04-24 end-to-end validation pass entry lists T0-T6; hot.md confirms all four mechanisms validated.
|
||||
|
||||
### H12
|
||||
- query: Why does the LLM Wiki pattern avoid re-querying and re-assembling the same sources every time?
|
||||
- correct: wiki/concepts/Query-Time Retrieval.md, wiki/concepts/Persistent Wiki Artifact.md, wiki/concepts/Source-First Synthesis.md
|
||||
- relevant: wiki/concepts/LLM Wiki Pattern.md, wiki/concepts/Compounding Knowledge.md
|
||||
- category: cross-page
|
||||
- rationale: Spans three pages: RAG baseline definition + durable memory object + synthesis discipline.
|
||||
|
||||
### H13
|
||||
- query: What are two independent ways to embed the LLM Wiki pattern in different editor environments?
|
||||
- correct: wiki/entities/Ar9av-obsidian-wiki.md, wiki/entities/Claudian-YishenTu.md
|
||||
- relevant: wiki/comparisons/claude-obsidian-ecosystem.md
|
||||
- category: cross-page
|
||||
- rationale: Ar9av setup.sh for multiple agents; Claudian native TypeScript plugin; both different deployment models.
|
||||
|
||||
### H14
|
||||
- query: How could you avoid re-processing unchanged sources when integrating community submissions?
|
||||
- correct: wiki/concepts/cherry-picks.md, wiki/concepts/Pro Hub Challenge.md
|
||||
- relevant: wiki/log.md
|
||||
- category: cross-page
|
||||
- rationale: Delta tracking manifest solves re-ingest problem; integration pattern in Pro Hub workflow.
|
||||
|
||||
### H15
|
||||
- query: What is the most direct way to add defuddle to claude-obsidian's current ingest pipeline?
|
||||
- correct: wiki/entities/kepano-obsidian-skills.md, wiki/concepts/cherry-picks.md
|
||||
- relevant: wiki/getting-started.md
|
||||
- category: cross-page
|
||||
- rationale: Cherry-pick #3 + getting-started ingest flow both discussed; requires both for full picture.
|
||||
|
||||
### H16
|
||||
- query: Where can I find Karpathy's original description of the wiki pattern?
|
||||
- correct: wiki/concepts/Persistent Wiki Artifact.md, wiki/concepts/Source-First Synthesis.md, wiki/entities/Andrej Karpathy.md
|
||||
- relevant: wiki/concepts/LLM Wiki Pattern.md
|
||||
- category: partial-recall
|
||||
- rationale: Multiple pages cite Karpathy gist; user remembers "gist" + "Karpathy."
|
||||
|
||||
### H17
|
||||
- query: What was inspired by dragon curves and what problem does it solve?
|
||||
- correct: wiki/concepts/DragonScale Memory.md
|
||||
- relevant: wiki/log.md
|
||||
- category: partial-recall
|
||||
- rationale: User recalls dragon analogy but not "DragonScale"; wants to understand fold operator.
|
||||
|
||||
### H18
|
||||
- query: What is the ~500-word file that gets updated at the end of every session?
|
||||
- correct: wiki/concepts/Hot Cache.md
|
||||
- relevant: wiki/getting-started.md, wiki/overview.md, wiki/hot.md
|
||||
- category: partial-recall
|
||||
- rationale: Exact description matches; user remembers fragment but not exact terminology.
|
||||
|
||||
### H19
|
||||
- query: Which community competition paid out Claude Credits and who won first?
|
||||
- correct: wiki/concepts/Pro Hub Challenge.md
|
||||
- relevant: wiki/entities/Claude SEO.md
|
||||
- category: partial-recall
|
||||
- rationale: User remembers community challenge and winners but not the name.
|
||||
|
||||
### H20
|
||||
- query: How many projects does the wiki track as competitors or adjacent to claude-obsidian?
|
||||
- correct: wiki/comparisons/claude-obsidian-ecosystem.md
|
||||
- relevant: wiki/concepts/cherry-picks.md
|
||||
- category: partial-recall
|
||||
- rationale: User remembers ecosystem matrix but not exact page; queries for count of projects.
|
||||
|
||||
### H21
|
||||
- query: How does the wiki pattern use blockchain or distributed ledger technology for vault consensus?
|
||||
- correct: null
|
||||
- relevant: wiki/concepts/DragonScale Memory.md
|
||||
- category: negative
|
||||
- rationale: Wiki does not discuss blockchain; DragonScale addresses are creation-order counter, not DLT.
|
||||
|
||||
### H22
|
||||
- query: What CRDT algorithm does the vault use for simultaneous multi-user editing?
|
||||
- correct: null
|
||||
- relevant: wiki/hot.md, wiki/concepts/DragonScale Memory.md
|
||||
- category: negative
|
||||
- rationale: No CRDT discussion; v1.7 mentions wiki-lock.sh for multi-writer safety (locking, not CRDTs).
|
||||
|
||||
### H23
|
||||
- query: Can I build an LLM wiki without installing Obsidian as a desktop application?
|
||||
- correct: null
|
||||
- relevant: wiki/concepts/LLM Wiki Pattern.md, wiki/getting-started.md
|
||||
- category: negative
|
||||
- rationale: Pattern is agnostic but this vault uses Obsidian. No Obsidian-free mode documented.
|
||||
|
||||
### H24
|
||||
- query: How do you customize the semantic embedding model weights for better tiling accuracy on your domain?
|
||||
- correct: null
|
||||
- relevant: wiki/concepts/DragonScale Memory.md, wiki/log.md
|
||||
- category: negative
|
||||
- rationale: Mechanism 3 uses locked ollama nomic-embed-text; no fine-tuning support documented.
|
||||
|
||||
### H25
|
||||
- query: Does the vault have CI/CD pipeline to automatically ingest new .raw files on every push?
|
||||
- correct: null
|
||||
- relevant: wiki/log.md, wiki/hot.md
|
||||
- category: negative
|
||||
- rationale: No GitHub Actions / CI-CD / automated ingest documented. Auto-commit hook exists but not scheduled CI.
|
||||
|
||||
---
|
||||
|
||||
## Benchmark Notes
|
||||
|
||||
**Coverage**: 47 wiki pages, ~50 queries → roughly 1 query per page on average; spread covers concepts, entities, comparisons, meta, folds, sources.
|
||||
|
||||
**Balance**: 25 derived (natural questions) + 5 synonym + 10 cross-page + 5 partial-recall + 5 negative.
|
||||
|
||||
**Targets** (rough expectation, not a gate):
|
||||
- v1.6 baseline top-1: ~50-65% on derived, much lower on hard
|
||||
- v1.7 hybrid top-1: ~75-90% on derived, ~50-70% on hard
|
||||
- Plan §7 ship gate (from v1.7 implementation plan): ≥30% reduction in "wrong page cited" errors. Verify by comparing the percentage-points improvement on top-1 across all 50 queries.
|
||||
@@ -0,0 +1,32 @@
|
||||
# Semantic Tiling Report
|
||||
|
||||
- generated: 2026-04-24T09:20:59Z
|
||||
- model: nomic-embed-text
|
||||
- ollama_url: http://127.0.0.1:11434
|
||||
- thresholds: error>=0.9, review=0.8-0.9
|
||||
- calibrated: False (using uncalibrated defaults)
|
||||
- pages scanned: 41; embedded: 21; skipped: 20
|
||||
- skipped reasons: embed_error=1, excluded filename=9, under wiki/meta/=10
|
||||
- cache hits: 0; recomputed: 21; orphans pruned: 0
|
||||
|
||||
## Errors (similarity >= 0.9)
|
||||
|
||||
- none
|
||||
|
||||
## Review (0.8 <= similarity < 0.9)
|
||||
|
||||
- `0.8822` wiki/concepts/Compounding Knowledge.md -- wiki/concepts/LLM Wiki Pattern.md
|
||||
- `0.8729` wiki/concepts/LLM Wiki Pattern.md -- wiki/questions/How does the LLM Wiki pattern work.md
|
||||
- `0.8681` wiki/comparisons/claude-obsidian-ecosystem.md -- wiki/sources/claude-obsidian-ecosystem-research.md
|
||||
- `0.8594` wiki/concepts/cherry-picks.md -- wiki/entities/Ar9av-obsidian-wiki.md
|
||||
- `0.8401` wiki/concepts/Compounding Knowledge.md -- wiki/questions/How does the LLM Wiki pattern work.md
|
||||
- `0.8400` wiki/entities/Claudian-YishenTu.md -- wiki/entities/Nexus-claudesidian-mcp.md
|
||||
- `0.8343` wiki/concepts/LLM Wiki Pattern.md -- wiki/entities/Andrej Karpathy.md
|
||||
- `0.8318` wiki/concepts/cherry-picks.md -- wiki/sources/claude-obsidian-ecosystem-research.md
|
||||
- `0.8244` wiki/comparisons/claude-obsidian-ecosystem.md -- wiki/concepts/cherry-picks.md
|
||||
- `0.8229` wiki/entities/kepano-obsidian-skills.md -- wiki/sources/claude-obsidian-ecosystem-research.md
|
||||
- `0.8171` wiki/entities/ballred-obsidian-claude-pkm.md -- wiki/sources/claude-obsidian-ecosystem-research.md
|
||||
- `0.8067` wiki/concepts/cherry-picks.md -- wiki/entities/kepano-obsidian-skills.md
|
||||
- `0.8064` wiki/entities/Ar9av-obsidian-wiki.md -- wiki/entities/kepano-obsidian-skills.md
|
||||
- `0.8052` wiki/concepts/LLM Wiki Pattern.md -- wiki/entities/rvk7895-llm-knowledge-bases.md
|
||||
- `0.8040` wiki/entities/Andrej Karpathy.md -- wiki/entities/rvk7895-llm-knowledge-bases.md
|
||||
Binary file not shown.
|
After Width: | Height: | Size: 585 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 285 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 227 KiB |
@@ -0,0 +1,70 @@
|
||||
---
|
||||
type: overview
|
||||
title: "Wiki Overview"
|
||||
created: 2026-04-07
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- meta
|
||||
- overview
|
||||
status: developing
|
||||
related:
|
||||
- "[[index]]"
|
||||
- "[[hot]]"
|
||||
- "[[log]]"
|
||||
- "[[dashboard]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
sources:
|
||||
---
|
||||
|
||||
# Wiki Overview
|
||||
|
||||
Navigation: [[index]] | [[hot]] | [[log]] | [[dashboard]]
|
||||
|
||||
---
|
||||
|
||||
## Purpose
|
||||
|
||||
This is the claude-obsidian demo vault. It demonstrates the [[LLM Wiki Pattern]] — a system for building persistent, compounding knowledge bases using Claude and Obsidian.
|
||||
|
||||
Run `/wiki` to scaffold this vault for your own domain and replace this overview.
|
||||
|
||||
---
|
||||
|
||||
## Current Seed Content
|
||||
|
||||
|
||||
**Concepts seeded:**
|
||||
- [[LLM Wiki Pattern]] — the core architecture
|
||||
- [[Hot Cache]] — session context mechanism
|
||||
- [[Compounding Knowledge]] — why the pattern works
|
||||
|
||||
**Entities seeded:**
|
||||
- [[Andrej Karpathy]] — originated the pattern
|
||||
|
||||
**Sources seeded:**
|
||||
- [[claude-obsidian-ecosystem-research]] — 16+ projects, 13 cherry-picks identified (2026-04-08)
|
||||
|
||||
---
|
||||
|
||||
## Current State
|
||||
|
||||
- Sources ingested: 2
|
||||
- Wiki pages: 26
|
||||
- Last activity: 2026-04-08 (v1.4.1 shipped, release session filed)
|
||||
|
||||
---
|
||||
|
||||
## Canvases
|
||||
|
||||
- [[claude-obsidian-presentation]] — Full presentation: hero, overview, skills, architecture, Wiki vs RAG, visual demos (2026-04-07)
|
||||
- [[AI Marketing Hub Cover Images Canvas]] — Cover image library for AI Marketing Hub brand assets
|
||||
|
||||
---
|
||||
|
||||
## Key Themes
|
||||
|
||||
**Knowledge compounds.** Unlike RAG, the wiki pre-compiles synthesis. Cross-references are already there. Contradictions are flagged. Every ingest enriches existing pages rather than adding isolated chunks.
|
||||
|
||||
**The hot cache is the force multiplier.** A ~500-word file captures recent context. New sessions start with full context at minimal token cost.
|
||||
|
||||
**Obsidian is the IDE, Claude is the programmer.** The graph view shows what's connected. The human curates sources and asks questions. Claude writes and maintains everything else.
|
||||
@@ -0,0 +1,52 @@
|
||||
---
|
||||
type: question
|
||||
title: "How does the LLM Wiki pattern work?"
|
||||
question: "How does the LLM Wiki pattern work and why is it better than RAG?"
|
||||
answer_quality: definitive
|
||||
created: 2026-04-07
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- question
|
||||
- llm-wiki
|
||||
- knowledge-management
|
||||
status: developing
|
||||
related:
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
- "[[Compounding Knowledge]]"
|
||||
- "[[Hot Cache]]"
|
||||
- "[[index]]"
|
||||
- "[[Wiki vs RAG]]"
|
||||
sources: []
|
||||
---
|
||||
|
||||
# How does the LLM Wiki pattern work?
|
||||
|
||||
**Question:** How does the LLM Wiki pattern work and why is it better than RAG?
|
||||
|
||||
## Answer
|
||||
|
||||
The [[LLM Wiki Pattern]] turns an LLM into a knowledge architect rather than a search engine.
|
||||
|
||||
**Standard RAG** (Retrieval-Augmented Generation): every query searches raw documents, retrieves chunks, and assembles an answer from scratch. Nothing is built up. Ask the same question twice — it does the same work twice.
|
||||
|
||||
**The wiki pattern** is different. When a source arrives, the LLM reads it and integrates it: updating entity pages, noting contradictions, adding cross-references. The synthesis is done once and persists. Every query benefits from all previous ingests.
|
||||
|
||||
### The three layers
|
||||
|
||||
1. **`.raw/`** — your source documents. Immutable. Claude reads, never modifies.
|
||||
2. **`wiki/`** — Claude-generated knowledge. Summaries, entities, concepts, synthesis.
|
||||
3. **`CLAUDE.md`** — the schema. Tells Claude how the wiki is structured and what to do.
|
||||
|
||||
### Why it compounds
|
||||
|
||||
See [[Compounding Knowledge]] for the full argument. The short version: each new source doesn't just add one page — it enriches 8-15 existing pages. The connections between pages are where the value lives, not the raw content itself.
|
||||
|
||||
### The hot cache shortcut
|
||||
|
||||
[[Hot Cache]] (wiki/hot.md) is a ~500-word summary of recent context. New sessions read it first. Cross-project references read it first. It prevents re-reading the whole wiki just to answer "where were we?"
|
||||
|
||||
(Source: [[LLM Wiki Pattern]])
|
||||
|
||||
## Confidence
|
||||
|
||||
definitive — this is the core concept the entire vault demonstrates.
|
||||
@@ -0,0 +1,71 @@
|
||||
---
|
||||
type: reference
|
||||
title: "Methodology Modes — Quick Decision Tree"
|
||||
tags:
|
||||
- reference
|
||||
- methodology
|
||||
- wiki-mode
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[methodology-modes-guide]]"
|
||||
- "[[wiki-mode]]"
|
||||
---
|
||||
|
||||
# Methodology Modes — Quick Decision Tree
|
||||
|
||||
Pick a vault organizational style with `bash bin/setup-mode.sh`.
|
||||
|
||||
```
|
||||
┌──────────────────────────────────┐
|
||||
│ Do you have a methodology │
|
||||
│ preference for organizing notes? │
|
||||
└──────────────┬───────────────────┘
|
||||
│
|
||||
┌────────────────┴────────────────┐
|
||||
NO YES
|
||||
│ │
|
||||
▼ ▼
|
||||
┌────────────────┐ ┌──────────────────────────────┐
|
||||
│ GENERIC │ │ How do you think about content│
|
||||
│ (v1.7 default; │ └────────────────┬──────────────┘
|
||||
│ no opinion) │ │
|
||||
└────────────────┘ ┌────────────────────┼────────────────────┐
|
||||
│ │ │
|
||||
▼ ▼ ▼
|
||||
"topic clusters "active projects "atomic claims
|
||||
+ navigation vs ongoing with IDs +
|
||||
by links" responsibilities" dense links"
|
||||
│ │ │
|
||||
▼ ▼ ▼
|
||||
┌──────────┐ ┌──────────┐ ┌────────────────┐
|
||||
│ LYT │ │ PARA │ │ ZETTELKASTEN │
|
||||
│ (MOCs) │ │ (4-bucket│ │ (timestamped │
|
||||
│ │ │ actionable) │ IDs, flat) │
|
||||
└──────────┘ └──────────┘ └────────────────┘
|
||||
```
|
||||
|
||||
## Quick comparison
|
||||
|
||||
| | Generic | LYT | PARA | Zettelkasten |
|
||||
|---|---|---|---|---|
|
||||
| Folders | sources/entities/concepts/sessions/ | mocs/notes/ | projects/areas/resources/archives/ | (none — flat) |
|
||||
| Navigation | folder browse | link follow via MOCs | folder browse | ID reference |
|
||||
| Naming | descriptive | descriptive | descriptive | timestamp prefix |
|
||||
| Best for | beginners, mixed use | knowledge clusters | knowledge workers, GTD | researchers, academics |
|
||||
| Discipline | low | medium | medium | high |
|
||||
|
||||
## Per-mode follow-up
|
||||
|
||||
After setting mode:
|
||||
|
||||
- **Generic**: nothing further; v1.7 behavior preserved
|
||||
- **LYT**: create your first MOC; new ingests update MOCs automatically via wiki-ingest
|
||||
- **PARA**: process the `wiki/projects/inbox/` + `wiki/resources/incoming/` buckets periodically
|
||||
- **Zettelkasten**: build the parent/child relationship graph as notes accumulate; rely on Obsidian Graph view
|
||||
|
||||
## Cross-reference
|
||||
|
||||
Full guide: [[methodology-modes-guide]]
|
||||
Skill: [[wiki-mode]] (`skills/wiki-mode/SKILL.md`)
|
||||
Router: `scripts/wiki-mode.py`
|
||||
Setup: `bash bin/setup-mode.sh`
|
||||
@@ -0,0 +1,106 @@
|
||||
---
|
||||
type: reference
|
||||
title: "Transport Fallback Decision Tree"
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[wiki-cli]]"
|
||||
- "[[mcp-setup]]"
|
||||
updated: 2026-05-17
|
||||
---
|
||||
|
||||
# Transport Fallback Decision Tree
|
||||
|
||||
claude-obsidian v1.7+ supports four ways to read and mutate vault notes. This document is the canonical decision tree skills consult when picking one.
|
||||
|
||||
## Quick reference
|
||||
|
||||
```
|
||||
┌─────────────────────────┐
|
||||
│ I need to read/write │
|
||||
│ a wiki note │
|
||||
└────────────┬────────────┘
|
||||
│
|
||||
▼
|
||||
┌─────────────────────────────────────────────┐
|
||||
│ Read .vault-meta/transport.json │
|
||||
│ (if missing, run scripts/detect-transport.sh)│
|
||||
└─────────────────────────────────────────────┘
|
||||
│
|
||||
┌────────────────┴────────────────┐
|
||||
│ │
|
||||
preferred = "cli" preferred = "filesystem"
|
||||
│ │
|
||||
▼ ▼
|
||||
┌──────────────────┐ ┌────────────────────┐
|
||||
│ obsidian-cli <op>│ │ Read/Write/Edit │
|
||||
│ │ │ tools with absolute│
|
||||
│ Falls through on │ │ vault paths │
|
||||
│ binary error to │ │ │
|
||||
│ filesystem. │ │ Final floor. │
|
||||
└──────────────────┘ └────────────────────┘
|
||||
```
|
||||
|
||||
## The fallback chain (highest to lowest precedence)
|
||||
|
||||
| # | Transport | When preferred | Strengths | Weaknesses |
|
||||
|---|-----------|----------------|-----------|------------|
|
||||
| 1 | **`cli`** | Desktop with Obsidian 1.12+ installed. Auto-detected by `scripts/detect-transport.sh`. | No MCP, no TLS, no plugins. Subprocess speed. Loud failure on missing binary. | Desktop-only. Single-vault per invocation. |
|
||||
| 2 | **`mcp-obsidian`** | When the user has the Local REST API plugin + an MCP server configured. Not auto-detected in v1.7 — user must edit `transport.json` by hand or override per call. | Persistent connection; rich tools (`mcp__obsidian-vault__*`); patch_note primitives. | Requires plugin install + TLS bypass for self-signed cert; reentrancy risk in self-MCP-calls. |
|
||||
| 3 | **`mcpvault`** | When user has `@bitbonsai/mcpvault` configured. Not auto-detected in v1.7. | No Obsidian plugin needed; BM25 search built in. | Filesystem-based — no Obsidian-native features (live render, hot reload). |
|
||||
| 4 | **`filesystem`** | Always available. Final floor. Used when nothing above is. | Zero setup. Works on any machine with the repo. | No Obsidian-aware features (no resolved Bases, no daily-note logic, no backlink graph). |
|
||||
|
||||
## When skills should consult this tree
|
||||
|
||||
Any skill that mutates a vault file should follow this pattern at the top:
|
||||
|
||||
```python
|
||||
# Pseudocode for any vault-mutating skill
|
||||
transport = read_json(".vault-meta/transport.json") # auto-created on first run
|
||||
match transport["preferred"]:
|
||||
case "cli":
|
||||
result = bash(f'obsidian-cli {operation} "{vault_root}" "{path}"')
|
||||
case "mcp-obsidian" | "mcpvault":
|
||||
result = mcp_call(operation, path)
|
||||
case "filesystem" | _:
|
||||
result = read_or_write_tool(absolute_path)
|
||||
```
|
||||
|
||||
Skills covered by this policy in v1.7:
|
||||
- `wiki-ingest` — page creation, frontmatter mutation, manifest updates
|
||||
- `wiki-query` — page reads (when standard or deep mode pulls full text)
|
||||
- `wiki-lint` — bulk reads for orphan/link analysis
|
||||
- `wiki-fold` — fold-page writes
|
||||
- `save` — session-note writes
|
||||
- `autoresearch` — source + concept page writes during the research loop
|
||||
|
||||
Skills that explicitly bypass the tree (always use direct filesystem):
|
||||
- `wiki-cli` — this skill IS the CLI wrapper; recursion is meaningless
|
||||
- `defuddle` — operates on HTTP-fetched content, not vault state
|
||||
- `obsidian-markdown` / `obsidian-bases` / `canvas` — reference skills, do not mutate the vault
|
||||
|
||||
## Refresh policy
|
||||
|
||||
`scripts/detect-transport.sh` rewrites `.vault-meta/transport.json` only if the existing snapshot is older than 7 days (`STALE_AFTER_DAYS=7`). To force a refresh after installing/uninstalling the Obsidian CLI:
|
||||
|
||||
```bash
|
||||
bash scripts/detect-transport.sh --force
|
||||
```
|
||||
|
||||
## Manual override
|
||||
|
||||
If you have an MCP transport configured (or any other transport not in the auto-detected set) and want claude-obsidian to use it as preferred:
|
||||
|
||||
1. Open `.vault-meta/transport.json` in any editor.
|
||||
2. Set `"preferred": "mcp-obsidian"` (or `"mcpvault"`, or any custom value).
|
||||
3. Set `"fallback_chain": ["mcp-obsidian", "filesystem"]` (or your preferred order).
|
||||
4. Add a `"manual_override": true` field at the top level so the detection script knows not to clobber your `preferred` + `fallback_chain` on next run.
|
||||
|
||||
The detection script (`scripts/detect-transport.sh`, v1.8.2+) parses the existing transport.json BEFORE running auto-detection. When `manual_override` is `true`, the script re-runs auto-detection only to refresh `available.cli.*` and `available.cli.obsidian_app_running` (so the human log line stays accurate), but preserves the user's `preferred` and `fallback_chain` choices across every cycle — including `--force` runs and post-staleness refreshes. The `manual_override` field itself round-trips through the snapshot, so the override survives indefinitely.
|
||||
|
||||
To disable the override later, edit the file and either set `"manual_override": false` or delete the field; the next detection pass will recompute `preferred` from auto-detection.
|
||||
|
||||
## See also
|
||||
|
||||
- Detection script: [`scripts/detect-transport.sh`](../../scripts/detect-transport.sh)
|
||||
- CLI recipe reference: [`skills/wiki-cli/SKILL.md`](../../skills/wiki-cli/SKILL.md)
|
||||
- Legacy 4-option MCP setup: [`skills/wiki/references/mcp-setup.md`](../../skills/wiki/references/mcp-setup.md) (still authoritative for MCP configuration steps)
|
||||
@@ -0,0 +1,42 @@
|
||||
---
|
||||
type: meta
|
||||
title: "Sources Index"
|
||||
updated: 2026-04-07
|
||||
tags:
|
||||
- meta
|
||||
- index
|
||||
- source
|
||||
status: evergreen
|
||||
related:
|
||||
- "[[index]]"
|
||||
- "[[log]]"
|
||||
- "[[entities/_index]]"
|
||||
- "[[Andrej Karpathy]]"
|
||||
---
|
||||
|
||||
# Sources Index
|
||||
|
||||
Navigation: [[index]] | [[concepts/_index|Concepts]] | [[entities/_index|Entities]]
|
||||
|
||||
All source pages — summaries of ingested documents, transcripts, articles, and data.
|
||||
|
||||
---
|
||||
|
||||
## Transcripts
|
||||
|
||||
|
||||
---
|
||||
|
||||
## Articles
|
||||
|
||||
<!-- Add article source pages here -->
|
||||
|
||||
---
|
||||
|
||||
## Papers
|
||||
|
||||
<!-- Add paper source pages here -->
|
||||
|
||||
---
|
||||
|
||||
## Add new sources here after each ingest.
|
||||
@@ -0,0 +1,53 @@
|
||||
---
|
||||
type: source
|
||||
title: "Claude + Obsidian Ecosystem Research"
|
||||
created: 2026-04-08
|
||||
updated: 2026-04-08
|
||||
tags:
|
||||
- research
|
||||
- ecosystem
|
||||
- competitive-analysis
|
||||
status: current
|
||||
related:
|
||||
- "[[claude-obsidian-ecosystem]]"
|
||||
- "[[cherry-picks]]"
|
||||
- "[[LLM Wiki Pattern]]"
|
||||
raw_file: ".raw/claude-obsidian-ecosystem-research.md"
|
||||
---
|
||||
|
||||
# Source: Claude + Obsidian Ecosystem Research
|
||||
|
||||
**Type**: Web research (GitHub + web search)
|
||||
**Date**: 2026-04-08
|
||||
**Queries run**: 6 parallel searches
|
||||
**Repos deep-read**: 12
|
||||
|
||||
## Summary
|
||||
|
||||
Comprehensive internet sweep of all Claude + Obsidian projects. Found 16+ active projects across 4 categories. Identified 13 specific features to cherry-pick for claude-obsidian v1.3.0+.
|
||||
|
||||
## Pages Created from This Source
|
||||
|
||||
- [[claude-obsidian-ecosystem]] — feature matrix comparison
|
||||
- [[cherry-picks]] — prioritized feature backlog
|
||||
- [[Ar9av-obsidian-wiki]] — entity page
|
||||
- [[Nexus-claudesidian-mcp]] — entity page
|
||||
- [[ballred-obsidian-claude-pkm]] — entity page
|
||||
- [[rvk7895-llm-knowledge-bases]] — entity page
|
||||
- [[kepano-obsidian-skills]] — entity page
|
||||
- [[Claudian-YishenTu]] — entity page
|
||||
|
||||
## Key Findings
|
||||
|
||||
1. **16+ projects** found combining Claude/AI with Obsidian
|
||||
2. **kepano** (Obsidian creator) validates Agent Skills format with his own skill repo
|
||||
3. **ballred/obsidian-claude-pkm** has best auto-commit implementation (PostToolUse hook)
|
||||
4. **Ar9av/obsidian-wiki** has best delta tracking (manifest.json)
|
||||
5. **rvk7895/llm-knowledge-bases** has best query depth system (3 tiers)
|
||||
6. Most popular traditional plugins: obsidian-copilot (5,776 ⭐), smart-connections (4,357 ⭐)
|
||||
7. **claude-obsidian unique advantages**: hot cache, canvas skill, /save conversation, marketplace polish
|
||||
8. **Top gap**: No URL ingestion — forces manual copy-paste from web
|
||||
|
||||
## Raw File
|
||||
|
||||
`.raw/claude-obsidian-ecosystem-research.md` — full notes with all sources
|
||||
Reference in New Issue
Block a user