Comprehensive assessment

Natural invoke

• "Audit the whole repo for template_active_inference, template_autoresearch_project, template_code_project, template_prose_project, and template_template" (example names only — the authoritative public roster is docs/_generated/active_projects.md)
• "Full health check — tests, docs, manuscript, pipeline"
• "What's broken across all active projects?"

Inputs to confirm

• Project(s) — names from docs/_generated/active_projects.md or "all active".
• Scope — single project vs multi-project (one pytest process per project test dir).

Workflow

Work through:

1. Tests — appropriate pytest invocations; one project test dir per project. Coverage: infra ≥60%, project ≥90% unless CI exception. No mocks (docs/rules/testing_standards.md).

1. Architecture — thin orchestrators in scripts/; algorithms in projects/<n>/src/ or infrastructure/. No cross-project imports.

1. Documentation — meaningful dirs have accurate AGENTS.md + README.md; links resolve; link _generated/active_projects.md, do not duplicate roster.

1. Manuscript — Pandoc-crossref + .bib → docs/guides/manuscript-semantics.md. Registry labels.yaml + [[FIG:]] → manuscript-cross-references. Validate markdown; PDF/logs if outputs exist.

1. Pipeline — core vs full DAG, optional LLM stages, copy to output/<name>/ per CLAUDE.md / RUN_GUIDE.

Deliverables

• Executive summary (pass/fail per area).
• Concrete issues: path, symptom, suggested fix.
• Commands run + raw exit status.
• Do not invent coverage percentages.

Verification commands

bash
uv sync
uv run pytest projects/<project>/tests/ --cov=projects/<project>/src --cov-fail-under=90 -q
uv run pytest tests/infra_tests/ --cov=infrastructure --cov-fail-under=60 -q
uv run python scripts/pipeline/stage_01_test.py --project <project>
uv run python -m infrastructure.validation.cli markdown projects/<project>/manuscript/
uv run python -m infrastructure.validation.cli prerender projects/<project>/manuscript --repo-root .

When NOT to use

• Single failing stage → pipeline-debugging
• Claim-by-claim manuscript pass → manuscript-claim-verification
• Determinism only → reproducibility-audit