FESADev/phases/1-linear-static-mitc4-rebaseline/step15.md

Sprint Contract: Phase 1 Evaluator Closeout

Objective

Perform an independent Phase 1 evaluation against the rebaseline contracts, docs, tests, reference artifacts, and validation commands.

Required Reading

• /AGENTS.md
• /PROGRESS.md
• /PLAN.md
• /docs/README.md
• /docs/HARNESS_ENGINEERING.md
• /docs/PRD.md
• /docs/ARCHITECTURE.md
• /docs/ADR.md
• /docs/NUMERICAL_CONVENTIONS.md
• /docs/ABAQUS_INPUT_SUBSET.md
• /docs/VERIFICATION_PLAN.md
• /docs/RESULTS_SCHEMA.md
• /docs/MITC4_FORMULATION.md
• /phases/1-linear-static-mitc4-rebaseline/index.json

Scope

• Review all completed rebaseline steps.
• Run validation.
• Produce pass/fail findings.
• Update phase status, PLAN.md, and PROGRESS.md with final Phase 1 state.

Allowed Files

• PLAN.md
• PROGRESS.md
• phases/index.json
• phases/1-linear-static-mitc4-rebaseline/index.json
• Optional evaluator feedback artifact under phases/1-linear-static-mitc4-rebaseline/

Explicit Non-Goals

• Do not implement missing solver behavior during evaluation.
• Do not approve unresolved reference gaps silently.
• Do not mark old superseded phase steps as evidence for new MITC4 formulation compliance unless they were revalidated.

Tests To Write First

• None. This is an evaluator sprint.

Reference Artifacts

• All accepted Phase 1 reference artifacts under references/.

Acceptance Commands

python scripts/validate_workspace.py

Evaluator Checklist

• All CRITICAL AGENTS.md rules are satisfied.
• TDD expectations are met for implementation steps.
• MITC4 implementation follows docs/MITC4_FORMULATION.md.
• Parser has not silently expanded unsupported Abaqus features.
• RF is recovered from full vectors.
• At least one stored displacement reference regression passes, and remaining PRD reference gaps are explicit.
• PLAN.md and PROGRESS.md are synchronized.

Handoff Requirements

• If passed, mark the rebaseline phase complete and move remaining Phase 2 or reference tasks to PLAN.md.
• If failed, write a concise feedback artifact with required fixes and commands to rerun.

Do Not

• Do not self-approve implementation without evidence from tests and reference comparisons.