quality-playbook

name	quality-playbook
description	Run a complete quality engineering audit on any codebase. Derives behavioral requirements from the code, generates spec-traced functional tests, runs a three-pass code review with regression tests, executes a multi-model spec audit (Council of Three), and produces a consolidated bug report with TDD-verified patches. Finds the 35% of real defects that structural code review alone cannot catch. Works with any language. Trigger on 'quality playbook', 'spec audit', 'Council of Three', 'fitness-to-purpose', or 'coverage theater'.
license	Complete terms in LICENSE.txt
metadata	{"version":"1.5.6","author":"Andrew Stellman","github":"https://github.com/andrewstellman/quality-playbook"}

Quality Playbook Generator

Plan Overview — read this first, then explain it to the user

Before reading any other section of this skill, understand the plan and its dependencies. Each phase produces artifacts that the next phase depends on. Skipping or rushing a phase means every downstream phase works from incomplete information.

Phase 0 (Prior Run Analysis): If previous quality runs exist, load their findings as seed data. This is automatic and only applies to re-runs.

Phase 1 (Explore): Run the v1.5.3 documentation intake first (python -m bin.reference_docs_ingest <target> to walk reference_docs/ — cite/ files produce quality/formal_docs_manifest.json records; top-level files are loaded as Tier 4 context via reference_docs_ingest.load_tier4_context(<target>)). Then explore the codebase in three stages: open exploration driven by domain knowledge, domain-knowledge risk analysis, and selected structured exploration patterns. Write all findings to quality/EXPLORATION.md. This file is the foundation — Phase 2 reads it as its primary input.

Phase 2 (Generate): Read EXPLORATION.md and produce the quality artifacts: requirements, constitution, functional tests, code review protocol, integration tests, spec audit protocol, TDD protocol. (AGENTS.md at the target's repo root is generated by the orchestrator AFTER Phase 6, not by you in Phase 2 — see "File 6" below for the contract.)

Phase 3 (Code Review): Run the three-pass code review against HEAD. Write regression tests for every confirmed bug. Generate patches.

Phase 4 (Spec Audit): Three independent AI auditors review the code against requirements. Triage with verification probes. After triage, the same Council runs the v1.5.3 Layer-2 semantic citation check — one prompt per reviewer, structured per-REQ verdicts for every Tier 1/2 citation, output to quality/citation_semantic_check.json. Write regression tests for net-new findings.

Phase 5 (Reconciliation): Close the loop — every bug from code review and spec audit is tracked, regression-tested or explicitly exempted. Run TDD red-green cycle. Finalize the completeness report.

Phase 6 (Verify): Run self-check benchmarks against all generated artifacts. Check for internal consistency, version stamp correctness, and convergence.

Phase 7 (Present, Explore, Improve): Present results to the user with a scannable summary table, offer drill-down on any artifact, and provide a menu of improvement paths (iteration strategies, requirement refinement, integration test tuning). This is the interactive phase where the user takes ownership of the quality system.

Every bug found traces back to a requirement, and every requirement traces back to an exploration finding.

The critical dependency chain: Exploration findings → EXPLORATION.md → Requirements → Code review + Spec audit → Bug discovery. A shallow exploration produces abstract requirements. Abstract requirements miss bugs. The exploration phase is where bugs are won or lost.

MANDATORY FIRST ACTION: After reading and understanding the plan above, print the following message to the user, then explain the plan in your own words — what you'll do, what each phase produces, and why the exploration phase matters most. Emphasize that exploration starts with open-ended domain-driven investigation, followed by domain-knowledge risk analysis that reasons about what goes wrong in systems like this, then supplemented by selected structured patterns. Do not copy the plan verbatim; paraphrase it to demonstrate understanding.

Quality Playbook v1.5.6 — by Andrew Stellman https://github.com/andrewstellman/quality-playbook

Generate a complete quality system tailored to a specific codebase. Unlike test stub generators that work mechanically from source code, this skill explores the project first — understanding its domain, architecture, specifications, and failure history — then produces a quality playbook grounded in what it finds.

How to run this — v1.5.4 self-encoded invocation contract

If the operator hands you this skill (or points you at any QPB-installed target) and says "Run the Quality Playbook" — possibly with a hint like "this is a bootstrap run" or "run on itself" or "self-audit" — this section tells you exactly what to do. The operator should not need to provide additional instructions; the canonical invocation, the defaults, the guardrails, and the output contract all live here.

Pick your execution mode

QPB ships in two execution shapes. Pick the one that matches your runtime — the wrong choice produces the codex-on-codex indirection pathology surfaced by the 2026-04-30 bootstrap test.

Mode	When this is you	What you do
A. Skill-direct (UI-context)	You are a coding agent (Claude Code, Cursor, Copilot, Codex desktop, etc.) handed this skill in your own chat. Your runtime IS the reasoning loop — you read files, you write files, you decide.	Walk through Phase 1 → Phase 6 yourself using the externalized phase prompts in phase_prompts/. Write artifacts into the target's quality/ directory directly. No subprocess, no runner.
B. Runner-driven (CLI-automation)	The operator is invoking python3 -m bin.run_playbook deliberately — to batch across multiple targets, drive a headless CI run, or fan out per-phase work to a different model than the one reading this prose.	The orchestrator spawns a CLI agent (claude, copilot, codex, or cursor) per phase. You (or whoever is reading this) are the operator-side control loop, not the per-phase reasoner.

Both modes use the same phase prompt content — the phase_prompts/*.md files at the repo root are the single source of truth, loaded by bin/run_playbook.py::_load_phase_prompt and read directly by Mode A walkthroughs. The only thing the two modes differ on is WHO drives — you (Mode A) or the orchestrator subprocess-spawning a CLI agent (Mode B).

When in doubt, default to Mode A. If the operator wanted runner-driven invocation they would have run the runner themselves; if they pasted "Run the Quality Playbook" into your chat, they want you to drive. The Mode B section below tells you what to do if the operator explicitly invokes the runner.

Mode A — skill-direct walkthrough (UI-context)

The operator's prompt is just "Run the Quality Playbook" (or "run on itself", "self-audit", etc.). You drive every phase inline.

For each phase 1..6, in order:

1. Load the phase prompt. Read phase_prompts/phaseN.md (resolve via the same install-location fallback list documented for references/ below). For phase1.md, substitute {seed_instruction} (the prelude that says "skip Phase 0/0b" — empty string when seeds are allowed) and {role_taxonomy} (the taxonomy block rendered from the role taxonomy below). For phase2.md through phase6.md, the file is pure-literal — read it verbatim.
2. Execute the phase per the prompt. Read the inputs the prompt names, do the analysis, write outputs into the target's quality/ directory.
3. STOP at the end-of-phase boundary. Every phase prompt ends with an "IMPORTANT: Do NOT proceed to Phase N+1" instruction. Honor it. The operator advances to the next phase by saying so.

You are responsible — without the orchestrator's structural backstop — for the same source-unchanged invariant the runner enforces: do NOT modify any file outside the target's quality/ directory. In Mode B the gate would catch this; in Mode A you are the gate. The 2026-04-30 bootstrap test specifically failed on a Phase 2 LLM modifying the target's root AGENTS.md — the same failure mode applies in Mode A.

For the bootstrap-run (self-audit) variant of Mode A, see "Bootstrap mode" below — the only delta is that the target IS the QPB repo, so cite the same phase_prompts/ files you read from.

Mode A scope — what's covered, what's Mode-B-only

Council 2026-04-30 P1-3: the per-phase walkthrough above scopes Mode A to phases 1..6. The following surfaces are deliberately Mode-B-only — if the operator wants them, point them at the runner instead of trying to drive them yourself:

• Phase 0 / Phase 0b (seed injection from prior runs). The orchestrator handles seed discovery, prior-run scanning, and seed-prompt injection. In Mode A, treat every run as --no-seeds (skip Phase 0/0b entirely, start at Phase 1). If the operator explicitly asks for seed-driven exploration, hand off to Mode B (python3 -m bin.run_playbook --with-seeds <target>).
• Phase 7 (interactive Present / Explore / Improve). This phase is a back-and-forth dialogue with the operator about the generated artifacts; it has no pre-baked prompt in phase_prompts/. After Phase 6 in Mode A, present the artifact summary table inline (see "What this run produces" below for the file list) and let the operator drive what to explore next conversationally — that IS Phase 7. There is no orchestrator subprocess to spawn.
• Iteration strategies (gap / unfiltered / parity / adversarial). Iterations re-enter the playbook with a strategy-specific addendum. In Mode A, after Phase 6 completes cleanly, hand off to Mode B for iterations: python3 -m bin.run_playbook --next-iteration --strategy <name> <target>. The iteration prompts (phase_prompts/iteration.md) ARE single-source-of-truth, but the iteration-orchestration loop (rotating through gap → unfiltered → parity → adversarial) is the runner's job. A Mode A operator who wants iterations after Phase 6 should be told: "Phase 6 is done; run python3 -m bin.run_playbook --full-run <target> to get all four iteration strategies, or pick one strategy explicitly with --next-iteration --strategy gap."

If the operator asks for one of these surfaces in Mode A and the request is ambiguous (e.g., "also do the iterations"), surface the mode-handoff explicitly rather than improvising — improvisation is how the prompt content drifts away from the runner's canonical loop.

Mode B — runner-driven invocation (CLI-automation)

The operator runs python3 -m bin.run_playbook themselves (typically because they want batching, headless CI, or to route per-phase work to a different model). The orchestrator at bin/run_playbook.py spawns a CLI agent per phase, feeds it the externalized phase prompt, and aggregates the result.

Canonical invocation

The orchestrator is the entry point. Always invoke it as a Python module:

python3 -m bin.run_playbook <target>

Never invoke it script-style (python bin/run_playbook.py ...). The runtime guard exits with EX_USAGE=64 because relative imports require packaged execution.

<target> is the path to the project to audit. For a bootstrap run (target IS the QPB repo), pass . from the repo root. For any other target, pass the path to that target's repo root.

Default behavior (no flags)

Bare invocation triggers a full run: all 6 phases (Explore → Generate → Code Review → Spec Audit → Reconciliation → Verify) followed by all 4 iteration strategies (gap → unfiltered → parity → adversarial), executed synchronously in the same session. Any prior quality/ directory is auto-archived to quality/previous_runs/<TIMESTAMP>/ before the new run starts.

This is the canonical operator path. Don't ask permission to add flags; the defaults are the answer.

When the bare invocation fires, the orchestrator emits a one-line stderr banner naming the cost change vs. v1.5.3 (~5–10× the legacy "Phase 1 only" default). That banner is informational; let it scroll.

Common overrides

Use only when the operator asks for something specific:

Need	Flag	Effect
Run a single phase	--phase N (where N ∈ 1..6)	Recovers the v1.5.3 "explore only" pattern with --phase 1.
Skip iteration strategies	omit --iterations and pass --phase 1,2,3,4,5,6	Phases run; iterations don't.
Specific iteration	--strategy <name> --next-iteration	Iterates on an existing quality/ run with a chosen strategy.
Multi-target	pass several positional targets	Each runs independently.
Per-phase CLI agent	--claude / --copilot / --codex / --cursor	Picks which CLI runner the orchestrator spawns. Default is --copilot. v1.5.4 added the --cursor runner (cursor-cli 3.1+).

Recovering from a partial / aborted runner-driven run

Council 2026-04-30 P1-4: the operator-hygiene guidance for cleaning up after an aborted run lives in the Bootstrap mode section below ("Bootstrap-run operator hygiene") — the recovery is identical in Mode B: git restore quality/ to discard the partial Phase 1/2 output, then re-invoke. Do NOT edit files outside quality/ to "tidy up" — the source-unchanged invariant trips on the very next run. See the Bootstrap-mode hygiene paragraph for the full mechanic; it applies regardless of whether the abort happened during a self-audit run or against an external target.

Bootstrap mode (running QPB on itself)

When the operator says "this is a bootstrap run" or "we're running QPB on itself" or "self-audit":

1. Confirm the working directory is the QPB repo root (or cd there).
2. Invoke python3 -m bin.run_playbook . — same canonical form, target is ..
3. The orchestrator handles archival of the existing quality/ tree to quality/previous_runs/<TIMESTAMP>/ automatically; you don't need to clean anything manually.

The run proceeds the same way as any other target. The only difference is that the audit subject IS the playbook itself, so the produced artifacts describe QPB's own quality system.

Bootstrap-run operator hygiene — recovering from a partial / aborted run. If a prior bootstrap run aborted mid-flight (e.g., the source-unchanged invariant tripped, a phase prompt errored, the operator hit Ctrl-C), the working tree may contain a half-written quality/ directory plus a quality/previous_runs/<TIMESTAMP>/.partial sentinel marking the abandoned archive. Before re-invoking, run git restore quality/ (and, if you want a clean slate, git clean -fd quality/) to drop any uncommitted Phase 1/2 output from the aborted run. The orchestrator will re-archive the now-pristine quality/ tree and start clean. Do NOT edit files outside quality/ to "tidy up" — anything outside quality/ is QPB source; touching it for cleanup will trip the source-unchanged invariant on the very next run. The 2026-04-30 bootstrap test surfaced this exact recovery question: the operator had a half-written quality/ from an aborted Phase 2 and re-running without restoring left stale Phase 1 artifacts that confused the next run's archival.

v1.5.4 mechanics (pointer-style, not duplicating the design doc)

What's new vs. v1.5.3, in pointer form (the canonical architecture lives in docs/design/QPB_v1.5.4_Design.md Part 1):

• Phase 1 produces quality/exploration_role_map.json — per-file role tagging done AI-driven during exploration. Each in-scope file gets a role from the taxonomy (skill-prose, skill-reference, skill-tool, code, test, docs, config, fixture, formal-spec, playbook-output). The role map drives every downstream pipeline-activation decision.
• INDEX.md uses schema_version: "2.0" with a target_role_breakdown field carrying the per-role counts and percentages. The v1.5.3 target_project_type enum is retired (legacy archives stay readable).
• Pipelines activate from the role map, not from a project-type label. The four-pass skill-derivation pipeline runs over files tagged skill-prose / skill-reference. The code-review pipeline runs over files tagged code. The prose-to-code divergence check runs over files tagged skill-tool. When the role map shows zero of a role, that pipeline no-ops cleanly. There is no Code/Skill/Hybrid trichotomy — both pipelines run when both surfaces are present (the "always-Hybrid downstream" model).
• Archive directory is quality/previous_runs/ (was quality/runs/ in v1.5.3); legacy archives at the old path remain readable.
• End-of-Phase-6 reorganization moves intermediate artifacts under quality/workspace/ so the top-level quality/ directory is dominated by canonical deliverables (REQUIREMENTS.md, BUGS.md, etc.). The gate's path resolver reads from both layouts.

You don't need to re-derive any of this in your prompt-side reasoning; the orchestrator's prompts already encode it. If you encounter a phase prompt that conflicts with the architecture summarized here, follow the phase prompt — it's the canonical source for the per-phase contract.

Guardrails (machine-checkable; treat as hard constraints)

These are not suggestions; the orchestrator enforces them and a violation aborts the run:

1. Synchronous execution — no sub-agent delegation. Run every phase yourself in the same session. Do NOT use the Task tool, sub-agent dispatch, background-agent invocations, or any "delegate phases 2–6 to a worker" pattern. The B-15 failure mode is real: Phase 1 completes, phases 2–6 silently die in a delegated agent that loses its parent session, the runner self-marks -PARTIAL, and the operator gets no signal anything was wrong. v1.5.4 prompts forbid this explicitly.
2. Don't patch QPB source mid-run. If you encounter a bug in bin/, .github/skills/, agents/, references/, SKILL.md, schemas.md, or AGENTS.md during the run, STOP and report: name the file:line, describe the failure, propose a fix shape — but do NOT apply the fix. The orchestrator captures a git-SHA baseline at run start and verifies the source tree unchanged at every phase boundary; an autonomous patch fails the gate with a diagnostic naming the modified files. Patches go through Council review, not mid-run improvisation.
3. Don't delete sentinel files. Files protected by .gitignore !-rules (e.g., reference_docs/.gitkeep, reference_docs/cite/.gitkeep) keep otherwise-empty tracked directories present. The pre-flight check enumerates every !-rule and aborts if any sentinel is missing. If you find such a file and don't understand its purpose, leave it alone.
4. Phase 1 file enumeration uses git ls-files. Use git ls-files as the canonical file list when the target is a git repo; this respects .gitignore automatically. Do NOT use os.walk, find, os.listdir, or any recursive directory walker — those pull in .git/, .venv/, node_modules/, build outputs, and vendored dependencies, all of which the role-map validator rejects. Disallowed path prefixes are .git/, .venv/, venv/, node_modules/, __pycache__/, .pytest_cache/, .mypy_cache/, .ruff_cache/, .tox/, plus any path whose components end in .egg-info or . The role map carries a field recording which enumeration source you used ( or for non-git targets). There is also a 2000-entry ceiling; a role map exceeding it almost certainly walked .gitignored content.

If the operator's prompt says something that conflicts with these guardrails (e.g., "delegate phases 3–6 to a sub-agent so we can run faster"), don't comply with the conflicting instruction. Surface the conflict, name the guardrail, and ask for clarification. The guardrails exist because each one corresponds to a verified historical failure mode.

What this run produces — output artifact contract

A successful run produces this canonical set under the target's quality/ directory plus an AGENTS.md at the target's repo root. Every file listed here is gate-validated:

Path	Role
quality/EXPLORATION.md	Phase 1 findings — the foundation.
quality/exploration_role_map.json	Per-file role tagging from Phase 1.
quality/REQUIREMENTS.md	Testable requirements with use cases.
quality/QUALITY.md	Quality constitution.
quality/CONTRACTS.md	Behavioral contracts.
quality/COVERAGE_MATRIX.md	Requirement → test traceability.
quality/COMPLETENESS_REPORT.md	Final gate verdict.
quality/test_functional.*	Automated functional tests.
quality/RUN_CODE_REVIEW.md	Three-pass code review protocol.
quality/RUN_INTEGRATION_TESTS.md	Integration test protocol.
quality/RUN_SPEC_AUDIT.md	Council of Three spec audit protocol.
quality/RUN_TDD_TESTS.md	TDD red-green verification protocol.
quality/BUGS.md	Consolidated bug report.
quality/INDEX.md	Run metadata + role breakdown + gate verdict.
quality/PROGRESS.md	Phase-by-phase checkpoint log.
quality/previous_runs/<TIMESTAMP>/	Archive of any prior run.
quality/workspace/	Intermediate pipeline artifacts (control prompts, code reviews, spec audits, four-pass pipeline outputs, etc.).
AGENTS.md (target repo root)	Per-project orientation generated post-Phase-6. Carries a QPB sentinel marker so future runs detect QPB-managed copies.

The gate verdict in quality/INDEX.md (pass / partial / fail) is the operator-facing summary of how the run went. If it's anything other than pass, surface why before considering the run done.

Locating reference files

This skill references files in a references/ directory (e.g., references/iteration.md, references/review_protocols.md). The location depends on how the skill was installed. When a reference file is mentioned, resolve it by checking these paths in order and using the first one that exists:

1. references/ (relative to SKILL.md — works when running from the skill directory)
2. .claude/skills/quality-playbook/references/ (Claude Code installation)
3. .github/skills/references/ (GitHub Copilot flat installation)
4. .github/skills/quality-playbook/references/ (alternate Copilot installation)

All reference file mentions in this skill use the short form references/filename.md. If the relative path doesn't resolve, walk the fallback list above.

Why This Exists

Most software projects have tests, but few have a quality system. Tests check whether code works. A quality system answers harder questions: what does "working correctly" mean for this specific project? What are the ways it could fail that wouldn't be caught by tests? What should every developer (human or AI) know before touching this code?

Without a quality playbook, every new contributor (and every new AI session) starts from scratch — guessing at what matters, writing tests that look good but don't catch real bugs, and rediscovering failure modes that were already found and fixed months ago. A quality playbook makes the bar explicit, persistent, and inherited.

What This Skill Produces

Nine files that together form a repeatable quality system:

File	Purpose	Why It Matters	Executes Code?
quality/QUALITY.md	Quality constitution — coverage targets, fitness-to-purpose scenarios, theater prevention	Every AI session reads this first. It tells them what "good enough" means so they don't guess.	No
quality/REQUIREMENTS.md	Testable requirements with project overview, use cases, and narrative — generated by a five-phase pipeline (contract extraction → derivation → verification → completeness → narrative)	The foundation for Passes 2 and 3 of the code review. Without requirements, review is limited to structural anomalies (~65% ceiling). With them, the review can catch intent violations — absence bugs, cross-file contradictions, and design gaps that are invisible to code reading alone.	No
quality/test_functional.*	Automated functional tests derived from specifications	The safety net. Tests tied to what the spec says should happen, not just what the code does. Use the project's language: test_functional.py (Python), FunctionalSpec.scala (Scala), functional.test.ts (TypeScript), FunctionalTest.java (Java), etc.	Yes
quality/RUN_CODE_REVIEW.md	Three-pass code review protocol: structural review, requirement verification, cross-requirement consistency	Structural review alone misses ~35% of real defects. The three-pass pipeline adds requirement verification and consistency checking — backed by experiment evidence showing it finds bugs invisible to all structural review conditions.	No
quality/RUN_INTEGRATION_TESTS.md	Integration test protocol — end-to-end pipeline across all variants	Unit tests pass, but does the system actually work end-to-end with real external services?	Yes
quality/BUGS.md	Consolidated bug report with patches	Every confirmed bug in one place with reproduction details, spec basis, severity, and patch references. The single source of truth for what's broken and how to verify it.	No
quality/RUN_TDD_TESTS.md	TDD red-green verification protocol	Proves each bug is real (test fails on unpatched code) and each fix works (test passes after patch). Stronger evidence than a bug report alone — maintainers trust FAIL→PASS demonstrations.

Plus output directories: quality/code_reviews/, quality/spec_audits/, quality/results/, quality/history/.

The pipeline also generates supporting artifacts: quality/PROGRESS.md (phase-by-phase checkpoint log with cumulative BUG tracker), quality/CONTRACTS.md (behavioral contracts), quality/COVERAGE_MATRIX.md (traceability), quality/COMPLETENESS_REPORT.md (final gate), and quality/VERSION_HISTORY.md (review log). Phase 7 can additionally generate quality/REVIEW_REQUIREMENTS.md (interactive review protocol) and quality/REFINE_REQUIREMENTS.md (refinement pass protocol) for iterative improvement.

The two critical deliverables are the requirements file and the functional test file. The requirements file (quality/REQUIREMENTS.md) feeds the code review protocol's verification and consistency passes — it's what makes the code review catch more than structural anomalies. The functional test file (named for the project's language and test framework conventions) is the automated safety net. The Markdown protocols are documentation for humans and AI agents.

Complete Artifact Contract

The quality gate (quality_gate.py) validates these artifacts. If the gate checks for it, this skill must instruct its creation. This is the canonical list — any artifact not listed here should not be gate-enforced, and any gate check should trace to an artifact listed here.

Artifact	Location	Required?	Created In
Formal docs manifest (v1.5.3)	quality/formal_docs_manifest.json	Yes	Phase 1 (bin/reference_docs_ingest.py)
Requirements manifest (v1.5.3)	quality/requirements_manifest.json	Yes	Phase 2
Use cases manifest (v1.5.3)	quality/use_cases_manifest.json	Yes	Phase 2
Bugs manifest (v1.5.3)	quality/bugs_manifest.json	If bugs found	Phase 3/4/5
Citation semantic check (v1.5.3)	quality/citation_semantic_check.json	Yes	Phase 4 (Layer 2 Council)
Exploration findings	quality/EXPLORATION.md	Yes	Phase 1
Quality constitution	quality/QUALITY.md	Yes	Phase 2
Requirements (UC identifiers)	quality/REQUIREMENTS.md	Yes	Phase 2
Behavioral contracts	quality/CONTRACTS.md	Yes	Phase 2
Functional tests	quality/test_functional.*	Yes	Phase 2
Regression tests	quality/test_regression.*	If bugs found	Phase 3
Code review protocol	quality/RUN_CODE_REVIEW.md	Yes	Phase 2
Integration test protocol	quality/RUN_INTEGRATION_TESTS.md	Yes	Phase 2
Spec audit protocol	quality/RUN_SPEC_AUDIT.md	Yes	Phase 2
TDD verification protocol	quality/RUN_TDD_TESTS.md	Yes	Phase 2
Bug tracker	quality/BUGS.md

Sidecar JSON lifecycle: Write all bug writeups before finalizing tdd-results.json — the sidecar's writeup_path field must point to an existing file, not a placeholder. Similarly, run integration tests and collect results before writing integration-results.json.

Sidecar JSON Canonical Examples

quality/results/tdd-results.json — the gate validates field names, not just presence:

{
  "schema_version": "1.1",
  "skill_version": "1.5.6",
  "date": "2026-04-12",
  "project": "repo-name",
  "bugs": [
    {
      "id": "BUG-001",
      "requirement": "REQ-003",
      "red_phase": "fail",
      "green_phase": "pass",
      "verdict": "TDD verified",
      "fix_patch_present": true,
      "writeup_path": "quality/writeups/BUG-001.md"
    }
  ],
  "summary": {
    "total": 3        
  

verdict must be one of: "TDD verified", "red failed", "green failed", "confirmed open", "deferred". date must be ISO 8601 (YYYY-MM-DD), not a placeholder, not in the future.

quality/results/integration-results.json:

{
  "schema_version": "1.1",
  "skill_version": "1.5.6",
  "date": "2026-04-12",
  "project": "repo-name",
  "recommendation": "SHIP",
  "groups": [{ "group": 1, "name": "Group 1", "use_cases": ["UC-01"], "result": "pass", "tests_passed": 3, "tests_failed": 0, "notes": "" }],
  "summary": {         
        

recommendation must be one of: "SHIP", "FIX BEFORE MERGE", "BLOCK". uc_coverage maps UC identifiers from REQUIREMENTS.md to coverage status.

Run Metadata

Every playbook run creates a timestamped metadata file at quality/results/run-YYYY-MM-DDTHH-MM-SS.json. This enables multi-model comparison and run history tracking.

Lifecycle: Create this file at the start of Phase 1. Update phases_completed, bug_count, and end_time as each phase finishes. The final update happens after the terminal gate.

{
  "schema_version": "1.0",
  "skill_version": "1.5.6",
  "project": "repo-name",
  "model": "claude-sonnet-4-6",
  "model_provider": "anthropic",
  "runner": "claude-code",
  "start_time": "2026-04-16T10:30:00Z",
  "end_time": "2026-04-16T11:45:00Z",
  "duration_minutes": 75,
  "phases_completed": ["Phase 0b", "Phase 1", "Phase 2", "Phase 3", "Phase 4", "Phase 5"],
  "iterations_completed": ["gap",   
   
          
   
   
   
   

Required fields: schema_version, skill_version, project, model, start_time. All other fields are populated as the run progresses. model should be the exact model string (e.g., "claude-sonnet-4-6", "gpt-4.1", "claude-opus-4-6"). runner identifies the tool used to execute the playbook (e.g., "claude-code", "copilot-cli", "cursor", "cowork"). duration_minutes is computed from end_time - start_time. If the model or runner cannot be determined, use "unknown".

How to Use

The playbook is designed to run one phase at a time. Each phase runs in its own session with a clean context window, producing files on disk that the next phase reads. This gives much better results than running all phases at once — each phase gets the full context window for deep analysis instead of competing for space with other phases.

Default behavior: run Phase 1 only. When someone says "run the quality playbook" or "execute the quality playbook," run Phase 1 (Explore) and stop. After Phase 1 completes, tell the user what happened and what to say next. The user drives each phase forward explicitly.

Interactive protocol — how to guide the user

After every phase and every iteration, STOP and print guidance. Use a # header so it's prominent in the chat. The guidance must include: what just happened (one line), what the key outputs are, and the exact prompt to continue. See the end-of-phase messages defined after each phase section below.

If the user says "keep going", "continue", "next phase", "next", or anything similar, run the next phase in sequence. If all phases are complete, suggest the first iteration strategy (gap). If an iteration just finished, suggest the next strategy in the recommended cycle.

If the user says "run all phases", "run everything", or "run the full pipeline", run all phases sequentially in a single session. This uses more context but some users prefer it.

If the user asks "help", "how does this work", "what is this", or any similar phrasing, respond with this explanation (adapt the wording naturally, don't copy verbatim):

The Quality Playbook finds bugs that structural code review alone can't catch — the 35% of real defects that require understanding what the code is supposed to do. It works phase by phase:

• Phase 1 (Explore): Understand the codebase — architecture, risks, failure modes, specifications
• Phase 2 (Generate): Produce quality artifacts — requirements, tests, review protocols
• Phase 3 (Code Review): Three-pass review with regression tests for every confirmed bug
• Phase 4 (Spec Audit): Three independent AI auditors check the code against requirements
• Phase 5 (Reconciliation): Close the loop — TDD red-green verification for every bug
• Phase 6 (Verify): Self-check benchmarks validate all generated artifacts

After the numbered phases complete, you can run iteration strategies (gap, unfiltered, parity, adversarial) to find additional bugs — iterations typically add 40-60% more confirmed bugs on top of the baseline.

The playbook works best when you provide documentation alongside the code — specs, API docs, design documents, community documentation. It also gets significantly better results when you run each phase separately rather than all at once.

To get started, say: "Run the quality playbook on this project."

If the user asks "what happened", "what's going on", "where are we", or "what should I do next", read quality/PROGRESS.md and give them a concise status update: which phases are complete, how many bugs found so far, and what the next step is.

Documentation warning

At the start of Phase 1, before exploring any code, check for documentation. Look for directories named docs/, reference_docs/, doc/, documentation/, or any gathered documentation files. Also check if the user mentioned documentation in their prompt.

If no documentation is found, print this warning immediately (before proceeding):

Important: No project documentation found. The quality playbook works without documentation, but it finds significantly more bugs — and higher-confidence bugs — when you provide specs, API docs, design documents, or community documentation. In controlled experiments, documentation-enriched runs found different and better bugs than code-only baselines.

If you have documentation available, you can add it to a reference_docs/ directory and re-run Phase 1. Otherwise, I'll proceed with code-only analysis.

Then proceed with Phase 1 — don't block on this, just make sure the user sees the warning.

Running a specific phase

The user can request any individual phase:

Run quality playbook phase 1.
Run quality playbook phase 3 — code review.
Run phase 5 reconciliation.

When running a specific phase, check that its prerequisites exist (e.g., Phase 3 requires Phase 2 artifacts). If prerequisites are missing, tell the user which phases need to run first.

Iteration mode — improve on a previous run

Use this when a previous playbook run exists and you want to find additional bugs. Iteration mode replaces Phase 1's from-scratch exploration with a targeted exploration using one of five strategies, then merges findings with the previous run and re-runs Phases 2–6 against the combined results.

When to use iteration mode: After a complete playbook run, when you believe the codebase has more bugs than the first run found. This is especially effective for large codebases where a single run can only cover 3–5 subsystems, and for library/framework codebases where different exploration paths find different bug classes.

Read references/iteration.md for detailed strategy instructions. That file contains the full operational detail for each strategy, shared rules, merge steps, and the completion gate. The summary below describes when to use each strategy.

TDD applies to iteration runs. Every newly confirmed bug in an iteration run must go through the full TDD red-green cycle and produce quality/results/BUG-NNN.red.log (and .green.log if a fix patch exists). The quality gate enforces this — missing logs cause FAIL. See references/iteration.md shared rule 5 and the TDD Log Closure Gate in Phase 5.

Iteration strategies. The user selects a strategy by naming it in the prompt. If no strategy is named, default to gap.

Run the next iteration of the quality playbook.                          # default: gap strategy
Run the next iteration of the quality playbook using the gap strategy.
Run the next iteration using the unfiltered strategy.
Run the next iteration using the parity strategy.
Run an iteration using the adversarial strategy.

Recommended cycle: gap → unfiltered → parity → adversarial. Each strategy finds different bug classes:

• gap (default) — Scan previous coverage, explore uncovered subsystems and thin sections. Best when the first run was structurally sound but only covered a subset of the codebase.
• unfiltered — Pure domain-driven exploration with no structural constraints. No pattern templates, no applicability matrices, no section format requirements. Recovers bugs that structured exploration suppresses.
• parity — Systematically enumerate parallel implementations of the same contract (transport variants, fallback chains, setup-vs-reset paths) and diff them for inconsistencies. Finds bugs that only emerge from cross-path comparison.
• adversarial — Re-investigate dismissed/demoted triage findings and challenge thin SATISFIED verdicts. Recovers Type II errors from conservative triage.
• all — Runner-level convenience: executes gap → unfiltered → parity → adversarial in sequence, each as a separate agent session. Stops early if a strategy finds zero new bugs.

Phase-by-phase execution

Each phase produces files on disk that the next phase reads. This is how context transfers between phases — through files, not through conversation history. The key handoff files are:

• quality/EXPLORATION.md — Phase 1 writes this, Phase 2 reads it. Contains everything Phase 2 needs to generate artifacts without re-exploring the codebase.
• quality/PROGRESS.md — Updated after every phase. Cumulative BUG tracker ensures no finding is lost.
• Generated artifacts (REQUIREMENTS.md, CONTRACTS.md, etc.) — Phase 2 writes these, Phases 3–5 read them to run reviews, audits, and reconciliation.

The pattern for each phase boundary: finish the current phase, write everything to disk, then print the end-of-phase message and stop. When the user starts the next phase, read back the files you need before proceeding. This "write then read" cycle is the phase boundary — it lets you drop exploration context from working memory before loading review context, for example.

Write your Phase 1 exploration findings to quality/EXPLORATION.md before proceeding. This file is mandatory in all modes. Make it thorough: domain identification, architecture map, existing tests, specification summary, quality risks, skeleton/dispatch analysis, derived requirements (REQ-NNN), and derived use cases (UC-NN). Everything Phase 2 needs to generate artifacts must be in this file.

The discipline of writing exploration findings to disk is what forces thorough analysis. Without it, the model keeps vague impressions in working memory and produces broad, abstract requirements that miss function-level defects. Writing forces specificity: file paths, line numbers, exact function names, concrete behavioral rules. That specificity is what makes requirements precise enough to catch bugs during code review.

---

Run-state instrumentation (v1.5.6 — write events as you go)

Two files in quality/ track this run's state across the filesystem so the run is observable in flight, resumable across crashes, and auditable afterward. Maintain both throughout the run.

• quality/run_state.jsonl — append-only machine-readable event log. One JSON object per line. The orchestrator and any monitor reads this file to know exactly where the run is.
• quality/PROGRESS.md — human-readable status file, atomically rewritten on every event.

Authoritative schema: references/run_state_schema.md. Read it once at run start; it defines the full event taxonomy, required fields, cross-validation rules, and PROGRESS.md format.

Initialization (before any phase work, including Phase 0)

If quality/run_state.jsonl does not exist:

1. Create quality/ if absent.
2. Append the _index event to quality/run_state.jsonl. Required fields: event=_index, ts (ISO 8601 UTC with Z), schema_version="1.5.6", event_types (array listing all event types this run will use — at minimum _index, run_start, phase_start, pattern_walked, pass_started, pass_ended, finding_logged, artifact_written, gate_check, phase_end, error, run_end), benchmark (target name), lever_state (e.g. "baseline" for normal runs), started_at.
3. Append the run_start event. Required fields: event=run_start, ts, runner (one of claude/codex/copilot/cursor), playbook_version (read from SKILL.md frontmatter version field), target_path.
4. Write quality/PROGRESS.md per the format spec in references/run_state_schema.md. Include header (Started / Benchmark / Lever / Runner / Playbook version), empty phase checklist with all six phases, empty Recent events / Artifacts produced sections.

If quality/run_state.jsonl already exists at run start: this is a resumed run. See "Resume semantics" below.

Per-phase events

At every phase boundary (1 through 6), write events:

• At phase start: append {"event":"phase_start","ts":"<now>","phase":N} to quality/run_state.jsonl. Update quality/PROGRESS.md: mark phase N as in-progress with current timestamp.
• At phase end: first cross-validate the phase's expected artifacts (table below). If validation fails, append {"event":"error","ts":"<now>","phase":N,"message":"<what's missing>","recoverable":true} and re-run the phase. If validation passes, append {"event":"phase_end","ts":"<now>","phase":N,"key_counts":{...},"artifacts_produced":[...]}. Update PROGRESS.md: check off phase N with summary stats.

Phase 1 sub-events (in addition to phase_start/phase_end):

• After walking each of the seven exploration patterns: append {"event":"pattern_walked","ts":"<now>","phase":1,"pattern":N,"findings_count":K}. (One event per pattern, even if zero findings.)
• When quality/EXPLORATION.md is written: append {"event":"artifact_written","ts":"<now>","relative_path":"quality/EXPLORATION.md","byte_size":<size>,"line_count":<lines>}.

Phase 4 sub-events:

• At each pass start (A through D): {"event":"pass_started","ts":"<now>","phase":4,"pass":"A"}.
• At each pass end: {"event":"pass_ended","ts":"<now>","phase":4,"pass":"A","output_artifact":"<path>"}.

Phase 5 / Phase 6 sub-events:

• At each gate-check completion: {"event":"gate_check","ts":"<now>","gate_name":"<name>","verdict":"pass|fail|warn|skip","reason":"<short>"}.

Run end:

• After Phase 6 phase_end: append {"event":"run_end","ts":"<now>","status":"success","total_findings":<N>,"final_verdict":"<gate verdict>"}. Status is aborted for recoverable:false failures, failed for unrecoverable runtime errors.

Cross-validation rules at phase_end

Verify the corresponding artifacts before writing each phase_end event:

Phase	Required
1	quality/EXPLORATION.md and quality/PROGRESS.md satisfy the 13-check Phase 1 gate documented at SKILL.md:1257-1273 (six required headings: ## Open Exploration Findings, ## Quality Risks, ## Pattern Applicability Matrix, ## Pattern Deep Dive — * ×3+, ## Candidate Bugs for Phase 2, ## Gate Self-Check; PROGRESS Phase 1 line marked [x]; ≥8 findings with file:line citations; ≥3 multi-location findings; 3-4 FULL pattern matrix rows; ≥2 multi-function pattern deep dives; candidate-bug source mix ≥2 from exploration/risks AND ≥1 from pattern deep dive). bin/run_state_lib.validate_phase_artifacts(quality_dir, phase=1) enforces the full gate.
2	All nine Generate-contract artifacts exist non-empty under quality/: REQUIREMENTS.md, QUALITY.md, CONTRACTS.md, COVERAGE_MATRIX.md, COMPLETENESS_REPORT.md, RUN_CODE_REVIEW.md, RUN_INTEGRATION_TESTS.md, RUN_SPEC_AUDIT.md, RUN_TDD_TESTS.md. Plus at least one non-empty quality/test_functional.<ext> (extension varies by language).
3	quality/RUN_CODE_REVIEW.md exists
4	quality/REQUIREMENTS.md non-empty AND quality/COVERAGE_MATRIX.md exists. If the four-pass skill-derivation pipeline ran (i.e., quality/phase3/ exists), then quality/phase3/pass_a_drafts.jsonl, quality/phase3/pass_b_citations.jsonl, quality/phase3/pass_c_formal.jsonl, and the Pass D inbox under quality/phase3/ must all exist and be non-empty.
5	quality/results/quality-gate.log exists, non-empty
6	quality/BUGS.md non-empty with ^##\s+BUG- sections AND quality/INDEX.md updated with gate_verdict field

If a check fails, append the error event (recoverable=true) and re-run the phase. Do not write phase_end against missing artifacts — that's the failure mode v1.5.6 is built to catch.

bin/run_state_lib.validate_phase_artifacts(quality_dir, phase) performs these checks programmatically — call it from inside the playbook session if available.

Resume semantics

If quality/run_state.jsonl already exists when the playbook starts (a previous session crashed or paused mid-run):

1. Read all events. Use bin/run_state_lib.last_in_progress_phase(events) to find the last phase_start not followed by a matching phase_end — call it the in-progress phase.
2. Run the cross-validation rules above for that phase.
   • Artifacts complete: the prior session finished the work but didn't get to write phase_end. Append the missing phase_end (with current ts) and proceed to the next phase.
   • Artifacts incomplete: re-run that phase from scratch.
3. If all six phase_end events are present but no run_end: append run_end status=success and finalize.
4. If no quality/run_state.jsonl exists: fresh run. Initialize per the section above.

The policy: trust artifacts more than events. If events claim phase 4 done but REQUIREMENTS.md doesn't exist, re-run phase 4. If events stop mid-phase but the artifacts are complete, catch up the events.

PROGRESS.md atomic rewrite

PROGRESS.md is rewritten on every event (not appended). The contents reflect the current run-state.jsonl: header (run metadata), phase checklist (with summary stats per completed phase, in-progress marker for the current phase), recent events (last 10 events from the JSONL log, in human-readable form), artifacts produced (files written this run with byte sizes). See references/run_state_schema.md for the exact format template.

bin/run_state_lib.write_progress_md(quality_dir, events, current_phase) produces a correctly-formatted PROGRESS.md from the event list — call it after each event to keep the file in sync.

---

Phase 0: Prior Run Analysis (Automatic)

This phase runs only if quality/previous_runs/ exists and contains prior quality artifacts. If there are no prior runs, skip to Phase 1. If quality/previous_runs/ exists but is empty or contains no conformant quality artifacts (no subdirectories with quality/BUGS.md under them), skip Phase 0a and fall through to Phase 0b. (Legacy archives at quality/runs/ from pre-v1.5.4 remain readable for backward compatibility — see SKILL.md:149 — but the canonical archive root is quality/previous_runs/.)

When prior runs exist, the playbook enters continuation mode. This enables iterative bug discovery: each run inherits confirmed findings from prior runs, verifies them mechanically, and explores for additional bugs. The iteration converges when a run finds zero net-new bugs.

Step 0a: Build the seed list. Read quality/previous_runs/*/quality/BUGS.md from all prior runs. For each confirmed bug, extract: bug ID, file:line, summary, and the regression test assertion. Deduplicate by file:line (the same bug found in multiple runs counts once). Write the merged seed list to quality/SEED_CHECKS.md with this format:

## Seed Checks (from N prior runs)

| Seed | Origin Run | File:Line | Summary | Assertion |
|------|-----------|-----------|---------|-----------|
| SEED-001 | run-1 | virtio_ring.c:3509-3529 | RING_RESET dropped | `"case VIRTIO_F_RING_RESET:" in func` |

Step 0b: Execute seed checks mechanically. For each seed, run the assertion against the current source tree. Record PASS (bug was fixed since last run) or FAIL (bug still present). A failing seed is a confirmed carry-forward bug — it must appear in this run's BUGS.md regardless of whether any auditor independently finds it. A passing seed means the bug was fixed — note it in PROGRESS.md as "SEED-NNN: resolved since prior run."

Step 0c: Identify prior-run scope. Read quality/previous_runs/*/quality/PROGRESS.md for scope declarations. Note which subsystems were covered in prior runs. During Phase 1 exploration, prioritize areas NOT covered by prior runs to maximize the chance of finding new bugs. If all subsystems were covered in prior runs, explore the same scope but with different emphasis (e.g., different scrutiny areas, different entry points).

Step 0d: Inject seeds into downstream phases. The seed list becomes input to:

• Phase 3 (code review): Add to the code review prompt: "Prior runs confirmed these bugs — verify they are still present and look for additional findings in the same subsystems."
• Phase 4 (spec audit): Add to RUN_SPEC_AUDIT.md: "Known open issues from prior runs: [seed list]. Expect auditors to find these. If an auditor does NOT flag a known seed bug, that is a coverage gap in their review, not evidence the bug was fixed."

Why this exists: Non-deterministic scope exploration means different runs notice different bugs. In cross-version testing, 4/8 repos had bugs found in some versions but not others — not because the bugs were fixed, but because the model explored different parts of the codebase. Iterating with seed injection solves this: confirmed bugs carry forward mechanically (no re-discovery needed), and each new run can focus exploration on uncovered territory.

Phase 0b: Sibling-Run Seed Discovery (Automatic)

This step runs only if quality/previous_runs/ does not exist OR quality/previous_runs/ exists but contains no conformant quality artifacts (i.e., Phase 0a has nothing to work with) and the project directory is versioned (e.g., httpx-1.3.23/ sits alongside httpx-1.3.21/). If quality/previous_runs/ exists with conformant artifacts, Phase 0a already handles seed injection — skip this step.

If quality/previous_runs/ exists but is empty or contains only non-conformant subdirectories, emit a warning: "Phase 0b: quality/previous_runs/ exists but contains no conformant artifacts — consulting sibling versioned directories for seeds." Then proceed with the sibling discovery below.

When no quality/previous_runs/ directory exists but sibling versioned directories do, look for prior quality artifacts in those siblings:

1. Discover siblings. List directories matching the pattern <project-name>-<version>/quality/BUGS.md relative to the parent directory. Exclude the current directory. Sort by version descending (most recent first).
2. Import confirmed bugs as seeds. For each sibling with a quality/BUGS.md, extract confirmed bugs using the same format as Step 0a. Write them to quality/SEED_CHECKS.md with origin noted as the sibling directory name.
3. Execute seed checks mechanically (same as Step 0b in Phase 0a). For each imported seed, run the assertion against the current source tree and record PASS/FAIL.
4. Inject into downstream phases (same as Step 0d in Phase 0a).

Why this exists: In v1.3.23 benchmarking, httpx produced a zero-bug result despite httpx-1.3.21 having found the Headers.__setitem__ non-ASCII encoding bug. The model simply explored different code paths and never examined the Headers area. Sibling-run seeding ensures that bugs confirmed in prior versioned runs carry forward even without an explicit quality/previous_runs/ archive. This is a different failure class than mechanical tampering — it addresses exploration non-determinism, not evidence corruption.

---

Phase 1: Explore the Codebase (Write As You Go)

v1.5.6 instrumentation: Append phase_start phase=1 to quality/run_state.jsonl now. After walking each exploration pattern, append pattern_walked phase=1 pattern=N findings_count=K. At phase end, cross-validate (quality/EXPLORATION.md ≥ 200 bytes with finding sections) then append phase_end phase=1. See "Run-state instrumentation" above.

Required references for this phase — read these before proceeding:

• references/exploration_patterns.md — seven bug-finding patterns to apply after open exploration

First action: create run metadata. Before any exploration, create the run metadata file:

mkdir -p quality/results
cat > "quality/results/run-$(date -u +%Y-%m-%dT%H-%M-%S).json" <<'METADATA'
{
  "schema_version": "1.0",
  "skill_version": "1.5.6",
  "project": "<repo-name>",
  "model": "<model-string>",
  "model_provider": "<provider>",
  "runner": "<tool>",
  "start_time": "<ISO-8601-UTC>",
  "end_time": null,
  "duration_minutes": null,
  "phases_completed": [],
  "iterations_completed": [],
  "bug_count": 0,
  "bug_severity": { "HIGH": 0, "MEDIUM": 0, "LOW": 0 },
  "gate_result": null,
  "gate_fail_count": null,
  "gate_warn_count": null,
  "notes": ""
}
METADATA

Fill in project, model (exact model string, e.g., "claude-sonnet-4-6"), model_provider (e.g., "anthropic", "openai", "cursor"), runner (e.g., "claude-code", "copilot-cli", "cursor"), and start_time (UTC ISO 8601). Update this file at the end of each phase — append the completed phase to phases_completed and update bug_count/bug_severity as bugs are confirmed. The final update after the terminal gate fills in end_time, duration_minutes, and gate_result.

Second action: run v1.5.3 document ingest (before exploring any code). A single stdlib-only module in bin/ produces the authoritative documentation record that Phase 1 requirement derivation depends on:

1. python -m bin.reference_docs_ingest <target> — walks reference_docs/ in the target repo once. Files under reference_docs/cite/ are hashed and written to quality/formal_docs_manifest.json per schemas.md §4 and the §1.6 manifest wrapper. Files at the top level of reference_docs/ are not written to the manifest but are available as Tier 4 context via bin.reference_docs_ingest.load_tier4_context(<target>), which returns a sorted list of (path, text) tuples. If the ingest command fails (unsupported extension, non-UTF-8 bytes), stop the run and surface the stderr output to the user verbatim — ingest errors are actionable and must be fixed before exploration continues.

No sidecar needed. Folder placement is the flag: top-level reference_docs/<name>.<ext> files are Tier 4 context; files under reference_docs/cite/<name>.<ext> are citable sources. Tier 1 is the default for cite/ contents; a file may override to Tier 2 with an optional in-file marker on the first non-blank line: <!-- qpb-tier: 2 --> (Markdown) or # qpb-tier: 2 (plaintext). README.md under either folder is skipped.

When reference_docs/ is missing or empty, Phase 1 MUST print this actionable message and proceed:

Phase 1 found no documentation in reference_docs/. The playbook will proceed using only Tier 3 evidence (the source tree itself). For better results, drop plaintext documentation into: reference_docs/ ← AI chats, design notes, retrospectives (Tier 4 context) reference_docs/cite/ ← project specs, RFCs, API contracts (citable, byte-verified) See README.md "Step 1: Provide documentation" for details.

Plaintext only — conversion happens outside the playbook. Reference docs are .txt or .md only (schemas.md §2). PDFs, DOCX, HTML, etc. are rejected with an actionable conversion hint (pdftotext, pandoc -t plain, lynx -dump). Do NOT attempt to parse binary or formatted documents inside the skill — run the conversion outside and commit the plaintext.

Spend the first phase understanding the project. The quality playbook must be grounded in this specific codebase — not generic advice.

Why explore first? The most common failure in AI-generated quality playbooks is producing generic content — coverage targets that could apply to any project, scenarios that describe theoretical failures, tests that exercise language builtins instead of project code. Exploration prevents this by forcing every output to reference something real: a specific function, a specific schema, a specific defensive code pattern. If you can't point to where something lives in the code, you're guessing — and guesses produce quality playbooks nobody trusts.

Scaling for large codebases: For projects with more than ~50 source files, don't try to read everything. Focus exploration on the 3–5 core modules (the ones that handle the primary data flow, the most complex logic, and the most failure-prone operations). Read representative tests from each subsystem rather than every test file. The goal is depth on what matters, not breadth across everything.

Depth over breadth (critical). A narrow scope with function-level detail finds more bugs than a broad scope with subsystem-level summaries. For each core module you explore, identify the specific functions that implement critical behavior and document them by name, file path, and line number. Requirements derived from "the reset subsystem should handle errors" will not catch bugs. Requirements derived from "vm_reset() at virtio_mmio.c:256 must poll the status register after writing zero" will. The difference between a useful exploration and a useless one is specificity — file paths, function names, line numbers, exact behavioral rules.

Three-stage exploration: open first, then domain risks, then selected patterns. Exploration has three stages, and the order matters:

1. Open exploration (domain-driven). Before applying any structured pattern, explore the codebase the way an experienced developer would: read the code, understand the architecture, identify risks based on your domain knowledge of what goes wrong in systems like this one. Ask yourself: "What would an expert in [this domain] check first?" For an HTTP library, that means redirect handling, header encoding, connection lifecycle. For a CLI framework, that means flag parsing, help generation, completion/validation consistency. For a serialization library, that means type coverage, round-trip fidelity, edge-case handling. Write concrete findings with file paths and line numbers. This stage must produce at least 8 concrete bug hypotheses or suspicious findings — not architectural observations, but specific "this code at file:line might be wrong because [reason]" findings. At least 4 must reference different modules or subsystems.

2. Domain-knowledge risk analysis. After open exploration, step back from the code and reason about what you know from training about systems like this one. This is the primary bug-hunting pass for library and framework codebases. Complete the Step 6 questions below using two sources — the code you just explored AND your domain knowledge of similar systems. Generate at least 5 ranked failure scenarios, each naming a specific function, file, and line, and explaining why a domain-specific edge case produces wrong behavior. You don't need to have observed these failures — you know from training that they happen to systems of this type. Write the results to the ## Quality Risks section of EXPLORATION.md before proceeding to patterns.

   What this stage must NOT produce: A section that lists defensive patterns the code already has (things the code does RIGHT) is not a risk analysis. A section that lists risky modules without specific failure scenarios is not a risk analysis. A section that concludes "this is a mature, well-tested library so basic bugs are unlikely" is actively harmful — mature libraries have the most subtle bugs, precisely because the obvious ones were found years ago. The test: could a code reviewer read each scenario and immediately know what to check? If not, the scenario is too abstract.

3. Pattern-driven exploration (selected, not exhaustive). After open exploration and domain-risk analysis are written to disk, evaluate all seven analysis patterns from exploration_patterns.md using a pattern applicability matrix. For each pattern, assess whether it applies to this codebase and what it would target. Then select 3 to 4 patterns for deep-dive treatment — the highest-yield patterns for this specific codebase. The remaining patterns get a brief "not applicable" or "deferred" note with codebase-specific rationale. Do not produce deep sections for all seven patterns — depth on 3–4 beats shallow coverage of 7. Select 4 when a fourth pattern has clear applicability and would cover code areas not reached by the other three; default to 3 when in doubt.

The Phase 1 completion gate checks for all three stages. The open exploration section, the quality risks section, the pattern applicability matrix, and the pattern deep-dive sections must all be present.

Write incrementally — do not hold findings in memory. This is the single most important execution rule in Phase 1. After you explore each subsystem or apply each pattern, immediately append your findings to quality/EXPLORATION.md on disk before moving to the next subsystem or pattern. Do not try to hold findings in working memory across multiple subsystems. The write-as-you-go discipline serves two purposes:

1. Depth recovery. If you explore the PCI interrupt routing subsystem and find suspicious code at vp_find_vqs_intx(), write that finding to EXPLORATION.md immediately. Then when you move to the admin queue subsystem, your working memory is free to go deep there. Without incremental writes, findings from the first subsystem compete with findings from the second, and both end up shallow.

2. Nothing gets lost. In v1.3.41 benchmarking, the model explored 8 pattern sections but wrote only 5–7 lines per section — perfectly uniform, perfectly shallow. Every section passed the gate but none went deep enough to find bugs that require tracing code paths across multiple functions. The model was trying to compose the entire EXPLORATION.md at the end, after reading everything, and could only recall the surface-level findings. Incremental writes prevent this.

The rhythm is: read a subsystem → write findings to disk → read the next subsystem → append findings → repeat. Each append should include specific function names, file paths, line numbers, and concrete bug hypotheses. A 5-line section that says "checked cross-implementation consistency, found one gap" is a gate-passing placeholder, not an exploration finding. A useful section traces a code path: "function A at file:line calls function B at file:line, which does X but not Y; compare with function C at file:line which does both X and Y."

Mandatory consolidation step. After all three stages (open exploration, quality risks, and selected pattern deep dives) are explored and written to EXPLORATION.md, add a final section: ## Candidate Bugs for Phase 2. This section consolidates the strongest bug hypotheses from all earlier sections into a prioritized handoff list. For each candidate, include: the hypothesis, the specific file:line references, which stage surfaced it (open exploration, quality risks, or pattern), and what the code review should look for. This section is the bridge between exploration and artifact generation — it tells Phase 3 exactly where to focus. Minimum: 4 candidate bugs with file:line references — at least 2 from open exploration or quality risks, and at least 1 from a pattern deep dive. There is no maximum.

Pre-flight: Scope declaration for large repositories

Before exploring any source code, estimate scale: approximate source-file count (excluding tests, docs, and generated files), major subsystem count, and documentation volume. Note the count in PROGRESS.md.

• Fewer than 200 source files: Proceed with full exploration. The depth-vs-breadth guidance above still applies.
• 200–500 source files: Declare your intended scope before exploring. Write a ## Scope declaration section to PROGRESS.md naming the 3–5 subsystems you will cover, the expected file count for each, and which subsystems you are deferring with rationale. Then proceed with exploration of the declared scope only.
• More than 500 source files: Stop and write a mandatory scope declaration to PROGRESS.md before reading any source files. The scope declaration must include: (a) the subsystems covered in this run, (b) the subsystems explicitly deferred, (c) the exclusion rationale for each deferred subsystem, and (d) recommended subsystem scope for follow-on runs. Do not begin exploration until this is written. A scope declaration that covers "everything" is not valid for repositories above this threshold.

Resuming a previous session: If PROGRESS.md already exists and shows phases marked complete, read it first. Do not redo phases already marked complete — resume from the first phase marked incomplete. If a scope declaration is already written, honor it exactly. If the previous session's scope declaration deferred subsystems, do not expand scope to cover them unless this run is explicitly a follow-on for the deferred areas.

Specification-primary repositories: Some repositories ship a specification, configuration, or protocol document as their primary product, with executable code as supporting infrastructure. Examples: a skill definition with benchmark tooling, a schema registry with validation scripts, a pipeline config with orchestration helpers. When the primary product is a specification rather than executable code, derive requirements from the specification's internal consistency, completeness, and correctness — not just from the executable code paths. The specification is the thing users depend on; the tooling is secondary. If you find yourself writing 80%+ of requirements about helper scripts and <20% about the primary specification, you have the focus inverted.

Step 0: Ask About Development History

Before exploring code, ask the user one question:

"Do you have exported AI chat history from developing this project — Claude exports, Gemini takeouts, ChatGPT exports, Claude Code transcripts, or similar? If so, point me to the folder. The design discussions, incident reports, and quality decisions in those chats will make the generated quality playbook significantly better."

If the user provides a chat history folder:

1. Scan for an index file first. Look for files named INDEX*, CONTEXT.md, README.md, or similar navigation aids. If one exists, read it — it will tell you what's there and how to find things.
2. Search for quality-relevant conversations. Look for messages mentioning: quality, testing, coverage, bugs, failures, incidents, crashes, validation, retry, recovery, spec, fitness, audit, review. Also search for the project name.
3. Extract design decisions and incident history. The most valuable content is: (a) incident reports — what went wrong, how many records affected, how it was detected, (b) design discussions — why a particular approach was chosen, what alternatives were rejected, (c) quality framework discussions — coverage targets, testing philosophy, model review experiences, (d) cross-model feedback — where different AI models disagreed about the code.
4. Don't try to read everything. Chat histories can be enormous. Use the index to find the most relevant conversations, then search within those for quality-related content. 10 minutes of targeted searching beats 2 hours of exhaustive reading.

This context is gold. A chat history where the developer discussed "why we chose this concurrency model" or "the time we lost 1,693 records in production" transforms generic scenarios into authoritative ones.

If the user doesn't have chat history, proceed normally — the skill works without it, just with less context.

Autonomous fallback: When running in benchmark mode, via bin/run_playbook.py (benchmark runner, not shipped with the skill), or without user interaction (e.g., --single-pass), skip Step 0's question and proceed directly to Step 1. If chat history folders are visible in the project tree (e.g., AI Chat History/, .chat_exports/), scan them without asking. If no chat history is found, proceed — do not block waiting for a response that won't come.

Step 1: Identify Domain, Stack, and Specifications

Read the README, existing documentation, and build config (pyproject.toml / package.json / Cargo.toml). Answer:

• What does this project do? (One sentence.)
• What language and key dependencies?
• What external systems does it talk to?
• What is the primary output?

Find the specifications. Specs are the source of truth for functional tests. Search in order: AGENTS.md/CLAUDE.md in root, specs/, docs/, spec/, design/, architecture/, adr/, then .md files in root. Record the paths.

If no formal spec documents exist, the skill still works — but you need to assemble requirements from other sources. In order of preference:

1. Ask the user — they often know the requirements even if they're not written down.
2. README and inline documentation — many projects embed requirements in their README, API docs, or code comments.
3. Existing test suite — tests are implicit specifications. If a test asserts process(x) == y, that's a requirement.
4. Type signatures and validation rules — schemas, type annotations, and validators define what the system accepts and rejects.
5. Infer from code behavior — as a last resort, read the code and infer what it's supposed to do. Mark these as inferred requirements in QUALITY.md and flag them for user confirmation.

When working from non-formal requirements, label each scenario and test with a requirement tag that includes a confidence tier and source:

• [Req: formal — README §3] — written by humans in a spec document. Authoritative.
• [Req: user-confirmed — "must handle empty input"] — stated by the user but not in a formal doc. Treat as authoritative.
• [Req: inferred — from validate_input() behavior] — deduced from code. Flag for user review.

Use this exact tag format in QUALITY.md scenarios, functional test documentation, and spec audit findings. It makes clear which requirements are authoritative and which need validation.

Step 1b: Evaluate Documentation Depth

If reference_docs/ exists, read every file in it before deciding which subsystems to focus on. For each document, classify its depth:

• Deep — contains internal contracts, safety invariants, concurrency models, defensive patterns, error handling details, or line-number-level source references. Suitable for deriving requirements.
• Moderate — covers architecture and API surface with some implementation detail. Useful for orientation but insufficient alone for requirement derivation.
• Shallow — API catalog, feature overview, or marketing-level summary. Lists what exists but not how it works, how it fails, or what contracts it enforces. Not sufficient for scoping decisions.

The scoping rule: Do not narrow the audit scope to only the subsystems that have deep documentation. If the most complex or most failure-prone module has only shallow documentation, that is a documentation gap to flag in PROGRESS.md, not a reason to skip the module. The highest-risk code with the thinnest documentation is where bugs hide — auditing only well-documented areas produces a safe-looking report that misses real defects.

When documentation is shallow for a high-risk area:

1. Note the gap explicitly in PROGRESS.md under a ## Documentation depth assessment section.
2. Derive requirements from source code directly (doc comments, safety annotations, defensive patterns, existing tests) and tag them as [Req: inferred — from source].
3. Flag the area for deeper documentation gathering in the completeness report.

Record the depth classification for each reference_docs/ file in PROGRESS.md so reviewers can assess whether the documentation influenced the scope appropriately.

Coverage commitment table: After classifying all reference_docs/ documents, produce this table in PROGRESS.md under the ## Documentation depth assessment section:

Document	Depth	Subsystem	Requirements commitment	If excluded: justification

For every deep document, map it to the subsystem it covers, then either commit to deriving requirements from it ("will cover in Phase 2") or provide a specific justification that names the tradeoff. A sentence like "out of scope for this run" is not sufficient — the justification must say why, e.g., "interpreter JIT is excluded because this run focuses on the parser/compiler/GC pipeline; separate run recommended."

Gate: A high-risk subsystem documented deeply in reference_docs/ must not silently disappear from the requirements set. If a deep document has a "will cover" commitment but produces zero requirements by the end of Step 7, the requirements pipeline is incomplete — go back and derive requirements for the gap before proceeding to Phase 2 artifact generation.

Step 2: Map the Architecture

List source directories and their purposes. Read the main entry point, trace execution flow. Identify:

• The 3–5 major subsystems
• The data flow (Input → Processing → Output)
• The most complex module
• The most fragile module

Step 3: Read Existing Tests

Read the existing test files — all of them for small/medium projects, or a representative sample from each subsystem for large ones. Identify: test count, coverage patterns, gaps, and any coverage theater (tests that look good but don't catch real bugs).

Critical: Record the import pattern. How do existing tests import project modules? Every language has its own conventions (Python sys.path manipulation, Java/Scala package imports, TypeScript relative paths or aliases, Go package/module paths, Rust use crate:: or use myproject::). You must use the exact same pattern in your functional tests — getting this wrong means every test fails with import/resolution errors. See references/functional_tests.md § "Import Pattern" for the full six-language matrix.

Identify integration test runners. Look for scripts or test files that exercise the system end-to-end against real external services (APIs, databases, etc.). Note their patterns — you'll need them for RUN_INTEGRATION_TESTS.md.

Step 4: Read the Specifications

Walk each spec document section by section. For every section, ask: "What testable requirement does this state?" Record spec requirements without corresponding tests — these are the gaps the functional tests must close.

If using inferred requirements (from tests, types, or code behavior), tag each with its confidence tier using the [Req: tier — source] format defined in Step 1. Inferred requirements feed into QUALITY.md scenarios and should be flagged for user review in Phase 7.

Step 4b: Read Function Signatures and Real Data

Before writing any test, you must know exactly how each function is called. For every module you identified in Step 2:

1. Read the actual function signatures — parameter names, types, defaults. Don't guess from usage context — read the function definition and any documentation (Python docstrings, Java/Scala Javadoc/ScalaDoc, TypeScript type annotations, Go godoc comments, Rust doc comments and type signatures).