agentleFS
Sign inStart free
r-irber-irbe/proof-skills/skills/lean-quality-engine/SKILL.md

lean-quality-engine

Skill3 starsChanged 24 days agoApache-2.0View on GitHub

USE FOR: assessing overall Lean 4 project quality across seven dimensions (soundness, faithfulness, completeness, novelty. DO NOT USE FOR: running a single enforcement script (@lean-enforcement). TRIGGERS: QA, quality gate, milestone check, project health, qa-engine.

The file

Create a free account to mount
Files
SKILL.md

lean-quality-engine

⚠️ MANDATORY (hot-tier): the soundness and faithfulness gates in §Behavioural rules are hard gates — a failure HALTS the QA lifecycle. Skipping Persist = incomplete.

Routing

  • USE FOR: running the seven-dimension quality lifecycle (Q1 soundness, Q2 faithfulness, Q3 completeness, Q4 novelty, Q5 elegance, Q6 integration, Q7 documentation); computing the weighted Q_score; deciding go/no-go at a milestone.
  • DO NOT USE FOR: running a single script (delegate to @lean-enforcement); reviewing one proof (delegate to @lean-proof-review); facilitating a council session (delegate to @lean-review-council).
  • TRIGGERS: QA, quality gate, milestone check, project health, qa-engine, quality score, QA lifecycle.

Workflow

  1. Discover [discover] — read prior Q_score, list active milestone, locate latest enforcement / council / coverage artefacts.
  2. Run hard gates [execute] — invoke @lean-enforcement for Q1 (axiom audit, sorry check, lean_diagnostics static abstract interpretation certificates: NatSubTruncation, NatDivModByZero, Zone DBM decreasing_by synthesis, Octagon conservation ±x_i ± x_j <= c, Bourdoncle WTO loop widening, trace partitioning) and Q2 (faithfulness via doc-requirements traceability and lean_counterexample_search with Galois-guided hypothesis domain pruning analyzeHeaderHypothesisDomains / pruneHeaderNatDomain).
  3. Halt-check [validate] — if Q1 or Q2 failed, STOP and route to recovery; do NOT proceed to soft gates.

Recovery & STOP

  • Q1 or Q2 fails → immediate STOP; escalate to human; do NOT compute Q_score.
  • Enforcement script crashes ×3 in one lifecycle → STOP, file a bug against @lean-enforcement.
  • Q_score regression ≥ 10 % vs prior milestone → STOP, ask before persisting.
  • Confidence < 90 % on dimension weighting → STOP, ask.

Handoffs

  • Predecessors / successors: see FM handoffs (grammar from ADR-0080).
  • Source spec: specs/lean/quality-engine/requirements.md — every G-rule traces to an AC there.
  • Related ADRs: ADR-0076, ADR-0080.
  • Reference: see references/lean-quality-engine-handbook.md for the seven-dimension matrix, weights, lifecycle phases, anti-patterns, and reliability-engineering details (original v1 content, preserved verbatim).

Detailed reference

Full encyclopedic reference, decision trees, and lookup tables live in references/lean-quality-engine-handbook.md .

See also

  • ../../references/lean-quality-engine-handbook.md -- Complete lean-quality-engine Handbook
  • ../lean-gateway/SKILL.md -- Top-level skill router

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free