Workflow Integration Test Green

[BLOCKING] Execute skill steps in declared order. NEVER skip, reorder, or merge steps without explicit user approval. [BLOCKING] Before each step or sub-skill call, update task tracking: set in_progress when step starts, set completed when step ends. [BLOCKING] Every completed/skipped step MUST include brief evidence or explicit skip reason. [BLOCKING] If Task tools are unavailable, create and maintain an equivalent step-by-step plan tracker with the same status transitions.

Quick Summary

Goal: [Workflow] Trigger the Integration Test Green workflow — run the WHOLE integration-test suite (or the named target), adjudicate every failure into a written Fault Verdict before any edit, fix at the owning layer, loop until a fresh full verify is green twice in a row with zero failures, then leave the spec TCs, the integration-test reference doc, and the feature docs in sync with the suite that actually exists.

Summary: Set the Goal Contract and explicit verification scope (WHOLE SYSTEM by default), then run /investigate → /integration-test-verify-loop → conditional /debug-investigate//fix → /spec [mode=sync] → /scan --target=integration-tests → /docs-update → /workflow-end → /watzup; every failing round requires a written Fault Verdict, owning-layer fix, inline /changes-review, Round Integrity Check, and fresh full re-verification until two consecutive zero-failure runs or bounded escalation.

• Testability contract: resolve Unit/Integration/System/E2E applicability from runner/config evidence; record owner/root/data, copy-ready full + focused commands, zero-match behavior, CI/simple-Windows entry, unique run/data identity, and repeat proof; unresolved applicable fields block handoff, while non-applicable tiers require evidence-backed N/A.

When to use: "make all integration tests pass", "fix the failing integration tests", "the suite is red after my change", "loop until all integration tests are green", "diagnose this flaky integration test". For AUTHORING new tests from specs use /workflow-write-integration-test; this workflow is for driving an EXISTING suite to green.

Workflow:

1. Detect — resolve the verification scope: the WHOLE system by default, or the target named in the prompt.
2. Converge — loop verify → adjudicate → fix → review the fix → re-verify until zero failures across 2 consecutive runs, bounded by a round cap and a Round Integrity Check.
3. Sync — reconcile spec TCs, the integration-test reference doc, and impacted feature docs with the final suite.

Key Rules:

• MUST ATTENTION default the verification scope to the WHOLE SYSTEM — every integration-test project via testProjectPattern > testProjects — and pass it EXPLICITLY to /integration-test-verify; never fall through to its change-scoped git auto-detect default.
• MUST ATTENTION adjudicate EVERY failure with /debug-investigate + /integration-test-review (report-only) into ONE written Fault Verdict — TEST-WRONG · TEST-NOT-OPTIMAL · SOURCE-WRONG · ENVIRONMENT-BLOCKED · AMBIGUOUS — with file:line evidence, BEFORE any edit.
• MUST ATTENTION fix at the lowest invariant-owning layer (Entity > Service > Handler), never the crash site; a SOURCE-WRONG fix KEEPS or STRENGTHENS the test that caught it.
• MUST ATTENTION run /changes-review (INLINE, report-only) on the fix diff of EVERY loop round that landed a fix, folding its validated findings back into that same round — no fix reaches the next round un-code-reviewed.
• MUST ATTENTION back every pass/fail claim with actual test-runner output (Passed/Failed/Skipped counts + failing names) — "all passed" without output is theater.
• MUST ATTENTION treat a shrinking executed-test count, a growing skipped count, or a narrowed scope as a REGRESSION → STOP and escalate; a suite that got greener by losing tests did not converge.
• MUST ATTENTION keep claims evidence-based (file:line) with confidence >80% to act.
• MUST ATTENTION keep task tracking updated as each step starts/completes.
• MUST ATTENTION define success criteria before execution (the Goal Contract is set FIRST) and loop until observable verification passes.
• MUST ATTENTION require integration tests to protect a named business rule/invariant and fail if that intent breaks.
• MUST ATTENTION arrange integration-test data through real use cases or valid seeded fixtures; never create impossible state through repository hacks.
• MUST ATTENTION verify integration suites with 2 consecutive passing runs without DB reset before declaring done.
• NEVER force green by weakening or removing assertions, adding skip annotations, widening assertion timeouts, wrapping a retry around a failing assertion, or narrowing the scope.
• NEVER skip mandatory workflow or skill gates.

IMPORTANT MANDATORY Steps: /investigate -> /integration-test-verify-loop -> /debug-investigate [on-failure] -> /fix [on-failure] -> /spec [mode=sync] -> /scan --target=integration-tests -> /docs-update -> /workflow-end -> /watzup

[BLOCKING] Step 0 — CREATE THE FULL TASK LIST BEFORE ANY VERIFICATION WORK. Call TaskList first (resume, never duplicate), then TaskCreate EVERY task below in one pass — before /investigate, before the first test run. A workflow that starts verifying with an empty task list has already lost the ability to show where it is, and an interrupted run cannot be resumed. — why: this loop can span many rounds and a context compaction mid-round; the task list is the only state that survives it.

Fixed tasks — created 1:1 from the canonical sequence in .claude/workflows.json, in order:

1. [1] investigate — resolve verification scope to a concrete project/suite list
2. [2] integration-test-verify-loop — drive the suite to green (parent of the per-round tasks)
3. [3] debug-investigate [on-failure] — traced root cause behind every Fault Verdict (CONDITIONAL)
4. [4] fix [on-failure] — resolve every verdict at the owning layer (CONDITIONAL)
5. [5] spec [mode=sync] — reconcile §8 TCs with the executing tests
6. [6] scan --target=integration-tests — regenerate the integration-test reference doc
7. [7] docs-update — update every other impacted doc
8. [8] workflow-end — close workflow state
9. [9] watzup — summarize the convergence trail
10. [10] final review — verify work quality + extract lessons (not a sequence step — the standing close-out task)

[3] and [4] are ROLL-UPS, not separate invocations. They appear in the canonical sequence so the conditional fix half is visible in the task list from the start, but /integration-test-verify-loop is their single executing owner: each firing happens INSIDE a round as [2.N.2] / [2.N.5] below. Complete [3]/[4] once the loop converges, summarizing which rounds fired them — or, if no round ever failed, complete them with the reason recorded (no failure in any round). NEVER run them a second time at workflow level after the loop returns — that would be a parallel fix loop the Inline Execution Gate forbids.

Per-round tasks (created when EACH round opens — round N is not planned until round N-1 reported):

• [2.N.1] verify — full run over {scope}, capture real counts (always)
• [2.N.2] debug-investigate — trace root cause of each failure (CONDITIONAL: only if round N reported failures — the round-N instance of [3])
• [2.N.3] integration-test-review — report-only fault gates (CONDITIONAL: same trigger)
• [2.N.4] fault verdict — one written verdict per failure (CONDITIONAL: same trigger)
• [2.N.5] fix — resolve at the owning layer (CONDITIONAL: same trigger — the round-N instance of [4])
• [2.N.6] changes-review — review this round's fix diff (CONDITIONAL: only if a fix landed)
• [2.N.7] round integrity check — counts not shrunk, scope not narrowed (always)

A conditional task whose trigger never fires is marked completed with the reason recorded (no failures this round), NEVER silently dropped — why: a skipped-and-unrecorded gate is indistinguishable from a forgotten one when someone audits the run later.

[BLOCKING] Each step MUST ATTENTION invoke its Skill tool — marking a task completed without skill invocation is a workflow violation. NEVER batch-complete validation gates.

[CRITICAL] Adjudicate-Before-Fix Gate: inside /integration-test-verify-loop, no edit may land before that failure has a written Fault Verdict backed by /debug-investigate's traced root cause AND /integration-test-review's gate findings. An unadjudicated failure gets "fixed" by whatever is nearest — which is almost always the assertion, and a weakened assertion protects nothing.

[CRITICAL] Per-Round Review Gate: the loop's only convergence signal is "the tests went green" — and a green test cannot see a fix made at the wrong layer, an invariant broken elsewhere, dead code, a leaked domain concept, or a security/performance regression. So EVERY round that lands a fix must run /changes-review (INLINE, report-only) over that round's fix diff, validate its findings, and resolve them in the SAME round. A round that leaves a validated review finding open has not finished, even if its tests are green. This subsumes the SOURCE-WRONG verdict's own changes-review obligation — once per round over the whole fix diff, never twice, and never as a nested review→fix loop.

[CRITICAL] Inline Execution Gate: /integration-test-verify-loop and the skills it drives (/integration-test-verify, /debug-investigate, /integration-test-review, /changes-review) run INLINE via the Skill tool — NEVER as sub-agents. /debug-investigate requires its /why-review gate in the SAME session/main agent, and /integration-test-review self-binds its own fix + re-review obligations; a sub-agent cannot own either or carry it back to the loop. Their OWN internal fan-outs (verify's per-project integration-tester agents, review's phase agents) remain sub-agents by their own design, so context stays bounded.

[CRITICAL] Documentation Sync Is Part Of Done: the loop deliberately defers ALL doc work while it churns, so steps 3–5 are not an optional tail. /spec [mode=sync] reconciles §8 TCs ↔ the executing test code, /scan --target=integration-tests regenerates the integration-test reference doc from the suite as it now stands, and /docs-update catches every other impacted doc. A converged-but-undocumented suite leaves the next agent reading a reference doc describing tests that no longer exist.

Goal Contract propagation (workflow-owned): At workflow start — BEFORE round 1 — resolve the active Goal Contract per SYNC:goal-contract-satisfaction-loop (active plan goal.md → plans/goals/{YYMMDD-HHmm}-{slug}/goal.md → create from the request). Its single required Success Criterion: a fresh full /integration-test-verify over the resolved scope reports ZERO failed tests across 2 consecutive runs without a DB reset, with no test deleted, skipped, or weakened to get there. Record the scope string, the round cap (default 3), and the baseline executed/skipped counts in Constraints. After every round, append the per-project counts, Fault Verdicts, and fixes to the Iteration Log; emit the Goal Satisfaction matrix (PASS/FAIL/BLOCKED) before /workflow-end.

Activate the workflow-integration-test-green workflow. Run /start-workflow workflow-integration-test-green with the user's prompt as context.

Test Architecture Contract Handoff

Before round 1, /investigate locks one evidence-backed contract record that /integration-test-verify-loop reuses on every round:

• applicability: mark the integration/system tier APPLICABLE only with verified runner/configuration evidence; record other tiers as N/A — <evidence> unless their configured owner is explicitly in scope.
• owner: keep /integration-test-verify-loop as convergence owner and name the existing conditional owners for diagnosis, review, fixing, and fix review.
• fullCommand and focusedCommand: bind the full command to the resolved whole-system scope and the focused command to an explicitly named target; both must be configured, copy-ready, fail invalid or zero-match selections, and expose a simple/Windows entry point when required.
• runIdentity and dataStrategy: use a unique non-sensitive run identity, valid public-use-case setup, explicit target/additive seed mode, and isolated mutable data for parallel workers.
• repeatProof and result: retain exact per-round counts, failing names, exit status, scope, and repeat/parallel evidence; a persistent-state scope needs two consecutive no-reset full runs.

The loop owns this handoff and the Round Integrity Check: /integration-test-verify receives the fixed scope and commands, while /debug-investigate, /integration-test-review, /fix, and /changes-review retain their existing conditional ownership and gates. The existing delegated order remains the only route; no fallback runner, narrowed scope, or destructive reset may replace missing evidence.

Steps: /investigate → /integration-test-verify-loop → /debug-investigate [on-failure] → /fix [on-failure] → /spec [mode=sync] → /scan --target=integration-tests → /docs-update → /workflow-end → /watzup

[CRITICAL] Recursive Until Green — the loop is the workflow, not a step inside it. Step 2 does NOT run once. It RECURSES: run the full suite → if ANY test fails, adjudicate and fix that failure → run the FULL suite again from scratch → repeat. The workflow is finished only when a fresh full run reports zero failures twice consecutively with no test deleted, skipped, or weakened. Reporting failures and stopping is NOT an outcome this workflow permits — either it converges, or it bounded-escalates via AskUserQuestion (round cap 3 · failures not shrinking across 2 rounds · failures increasing · coverage lost · an open validated review finding · ENVIRONMENT-BLOCKED). — why: "the tests fail" is a status, not a deliverable; the caller asked for a green suite, and a workflow that hands back a red one has done the diagnosis and skipped the job.

/debug-investigate and /fix are CONDITIONAL steps of the loop, executed INSIDE /integration-test-verify-loop. They fire on every round that reports a failure and are skipped (with a recorded reason) on a round that is already green. They are tracked as their own tasks per round (Step 0) so they are visible in the task list, but /integration-test-verify-loop remains their single executing owner — NEVER invoke them as a second, parallel fix loop at workflow level. — why: the fix half was previously triple-owned across three skills, which let two loops double-fix one failure or each assume the other owned it; one owner with visible sub-tasks keeps both the accountability and the visibility.

Step 2 RECURSES: /integration-test-verify → (on failure) /debug-investigate → /integration-test-review → Fault Verdict → /fix → /changes-review → fresh full re-verify — repeating until the suite is green twice consecutively. Steps 3 and 4 of the sequence are the workflow-level roll-ups of that conditional half; /integration-test-verify-loop executes them inside each round, never again after it returns.

[STEP PURPOSES] Every step has a distinct purpose — NEVER deduplicate or batch:

/investigate — Resolve the verification scope to a concrete test-project list. No target in the prompt → the WHOLE system (every project via testProjectPattern > testProjects from docs/project-config.json → integrationTestVerify). A named suite/module/feature/diff narrows it — state how the target maps to projects. Output: the fixed scope string the loop will reuse every round. /integration-test-verify-loop — The convergence engine, and the only step that changes code. Sets the Goal Contract first, then loops: /integration-test-verify INLINE over the fixed scope (2 consecutive green runs, no DB reset, real counts) → on ANY failure run /debug-investigate + /integration-test-review (report-only) → ONE Fault Verdict per failure → /fix at the owning layer → conditional /changes-review (INLINE, report-only, over the round's fix diff — runs in EVERY round that landed a fix; validated findings fold back into that same round's fix set) → Round Integrity Check (executed count must not shrink, skipped count must not grow, scope must not narrow) → fresh full re-verify. Round cap 3; not shrinking across 2 rounds, increasing failures, cap hit with failures open, lost coverage, an open validated review finding, or ENVIRONMENT-BLOCKED → STOP and escalate via AskUserQuestion. Output: zero-failure runner evidence for both runs + the per-round verdict/fix/review trail. /debug-investigate (CONDITIONAL — only on a round with failures) — Trace each failure end-to-start to its root cause BEFORE any edit; produces the traced cause that the Fault Verdict rests on. Skipped on a green round, with the skip recorded. /fix (CONDITIONAL — only on a round with an adjudicated failure) — Resolve the verdict at the lowest invariant-owning layer (Entity > Service > Handler), never the crash site. A SOURCE-WRONG fix KEEPS or STRENGTHENS the test that caught it. NEVER runs before a written Fault Verdict exists for that failure. /spec [mode=sync] — Reconcile §8 TC-{FEATURE}-{NNN} specs ↔ the executing test code under docs/specs/. Update each TC's CoveredBy field with all covering {File}::{MethodName} links (one TC → many tests, 1:N; a test-filter expression when the set is large). Coverage = ≥1 annotation-tagged test; never force one test per TC. Runs AFTER convergence so it syncs the final tests, not intermediate ones. /scan --target=integration-tests — Regenerate the integration-test project-reference doc from the suite as it now stands: patterns, base fixtures, async-wait and unique-data helper conventions, suite/project inventory, and lessons. This is the doc every future agent reads before touching a test — a loop that changed test structure without regenerating it leaves the next agent following stale conventions. /docs-update — Update every OTHER impacted doc: feature-doc evidence fields, version history, and any doc embedding test counts or coverage claims the loop changed. Covers what /spec [mode=sync] (spec TCs) and /scan --target=integration-tests (the reference doc) do not. /workflow-end + /watzup — Close workflow state, then summarize the convergence trail and run the final handoff.

IMPORTANT MANDATORY Steps: /investigate -> /integration-test-verify-loop -> /debug-investigate [on-failure] -> /fix [on-failure] -> /spec [mode=sync] -> /scan --target=integration-tests -> /docs-update -> /workflow-end -> /watzup

Integration Test Execution Discipline — How the integration-test family (write · review · verify) runs, diagnoses, and clears a suite. Binds /integration-test, /integration-test-review, and /integration-test-verify identically.

1. Verify the WHOLE system passes — not a hand-picked subset. /integration-test-verify must prove the full relevant suite is green (every test in the system the change can touch), not one cherry-picked test. "All pass" is only true with actual runner output (Passed/Failed/Skipped counts + names) and only after 2 consecutive green runs without a DB reset.
2. Drive state through real use-case paths — NEVER hack seed data. Set up every precondition exactly as a real user would: real commands, queries, production consumers/messages, or valid idempotent seeders. NEVER create or mutate domain data by direct repository writes — that fabricates states a user could never reach and hides the real workflow bug. Hacking seed data to force a green run is forbidden.
3. On ANY failure → /debug-investigate the root cause BEFORE any fix. Do not guess, do not patch the symptom site. Trace the failure end-to-start and classify whose fault it is: test code (wrong assertion/setup), source/production code (real defect), or environment/infrastructure/data. Then route: test-code fault → /integration-test-review to fix the test at the root (never weaken assertions or add skips); source-code fault → fix the production defect at the owning layer and report it; environment fault → mark BLOCKED and point at the startup script. NEVER change a test to match broken code.
4. 60-second runtime cap — a slow test is a RED FLAG, not a tuning knob. Local integration tests run fast. If any single test (or a stalled suite) exceeds ~60s, STOP and treat the slowness itself as a defect signal — deadlock, missing await, infinite poll/retry, a real network/external call, or an unbounded query. /debug-investigate the cause; NEVER paper over it by raising the timeout or extending the wait.
5. Loop until the whole suite is green. After fixing the validated root cause, restart the full 2-run verification from run 1. Done means the entire relevant suite passes repeatably — never green-once, never a subset.

Test-Failure Fault Adjudication — When a test fails (or you are debugging or fixing a failure), the job is to determine who is at fault — the source code or the test code. Getting that verdict right matters more than turning the suite green. Binds every debug / fix / test skill identically.

1. Provisional verdict before touching either side. Classify the observed evidence as SOURCE-WRONG, TEST-WRONG, TEST-NOT-OPTIMAL, ENVIRONMENT-BLOCKED, or AMBIGUOUS; then /debug-investigate and trace end-to-start before editing. A green-again suite is NOT the goal.
2. Triangulate against the spec AND the source. If a governing Feature Spec covers the behavior (e.g. docs/specs/** — §3 ACs / §4 BRs / §5 invariants / §8 TCs), it is the tiebreaker for intended behavior — compare BOTH the production source and the failing test against it. With no spec, the documented intent / acceptance criteria / caller contract is the reference. Decide from this evidence whether the SOURCE is wrong or the TEST is wrong.
3. Classify who is at fault, then fix the wrong side at its root:
   • SOURCE-WRONG — production code violates the spec's intended behavior or a clear invariant → fix the source at the owning layer; keep or strengthen the test that caught it.
   • TEST-WRONG — the test encodes a stale or incorrect assertion, setup, or expectation that contradicts intended behavior → fix the test at its root. NEVER weaken an assertion, add a skip, or relax a timeout to force green.
   • TEST-NOT-OPTIMAL — intended behavior is valid but the test seam, timing, or assertion signal is fragile → improve the test without weakening the invariant.
   • ENVIRONMENT-BLOCKED — infrastructure or external state prevents a source/test verdict → preserve diagnostics and stop mutation until the environment is healthy.
   • AMBIGUOUS — evidence or intended behavior does not safely select an owner → ask the user or canonical owner before editing.
   • NEVER change a test to match broken source, and NEVER change source to satisfy a broken test. (Migration code excluded — schema/data migrations are one-time execution paths, not core application logic.)
4. Ask the user when intended behavior is unclear. If no spec covers the behavior, the spec is silent, or the spec is ambiguous about which side is correct, STOP and AskUserQuestion (or consult the canonical spec owner) before editing either side — never silently pick source or test just to make the suite pass.

Reconcile to intended behavior, never to whichever side currently passes — green can encode the very bug.

Read-only/report-only role boundary: when this block is carried by a report-only role (code-reviewer, quality-gate-review, spec-compliance-reviewer, tester, and any other agent whose definition declares it never edits source), "fix the wrong side" means RETURN the adjudicated verdict and the proposed repair to the parent — do not modify source, tests, generated carriers, or user data. The adjudication is the deliverable; the edit is the caller's. Without this sentence the block's step-3 imperatives read as write authority and directly contradict those agents' own declarations (e.g. tester.md "NEVER implement fixes"), which is the sibling SYNC:double-round-trip-review boundary applied to the same class of carrier.

Real-World Fidelity Gate — MANDATORY when authoring, reviewing, or repairing any integration / E2E / system test.

A test earns trust by reproducing a situation the system can actually meet in production. A scenario that could never occur in real life proves nothing when it passes, and wastes hours when it fails.

1. Ask the fidelity question BEFORE writing the setup: "Can this sequence, timing, and data actually occur in production?" If no, the test is mis-specified — fix the SCENARIO, never the assertion.
2. Model real pacing between actor steps. Two distinct actor actions that production separates by seconds, minutes, or hours MUST NOT be fired back-to-back in the same millisecond. Compressed pacing manufactures races the system was never designed to survive, then reports them as product defects.
3. Wait on a real signal, never a blind sleep. Find an observable proving the prior step finished — a persisted state change, an audit/version stamp, a queue/worker idle marker, a completion event — and poll until it settles (unchanged across a short stability window). Use a fixed delay ONLY when no observable exists, and say so in a comment.
4. Barriers belong in ARRANGE, never in ASSERT. Waiting for a precondition is fidelity. Widening an assertion's timeout, loosening a comparison, adding a retry around a failing assertion, or skipping the test is masking. NEVER do the latter to force green.
5. Distinguish harness-amplified from real. Test topologies (shared infra, fan-out consumers, parallel suites, cold starts) can make a rare production race routine locally. Before filing a product defect, state whether the trigger exists in production and at what likelihood.
6. Keep the protected invariant intact. Improving fidelity must NEVER reduce what the test protects. If a realistic scenario no longer exercises the rule, the rule needs a DIFFERENT realistic scenario — not a weaker assertion.
7. Deliberate impossible-state tests are allowed, but MUST be labelled. Corruption-repair, migration, and fail-safe tests intentionally construct states production should never reach; comment WHY the state is reachable (upstream bug, partial write, legacy data), so they are never confused with unrealistic setups.

AI Mistake Prevention — Failure modes to avoid on every task:

Re-read files after context changes. Context compaction, resume, or long-running work can make memory stale; verify current files before acting. Verify generated content against source evidence. AI hallucinates APIs, names, claims, and document facts. Check the relevant source before documenting or referencing. Check downstream references before deleting or renaming. Removing an artifact can stale docs, generated mirrors, configs, and callers; map references first. Trace the full impact chain after edits. Changing a definition can miss derived outputs and consumers. Follow the affected chain before declaring done. Verify ALL affected outputs, not just the first. One green check is not all green checks; validate every output surface the change can affect. Assume existing values are intentional — ask WHY before changing OR flagging one as a defect. Before changing or reporting a constant, limit, flag, cutoff, wording, or pattern, read nearby context and history, the CALLER's ordering, and 2+ sibling call sites of the same convention. A doc stating WHAT without WHY is missing rationale, not proof of a missing guard. Surface ambiguity before acting — don't pick silently. Multiple valid interpretations require an explicit question or stated assumption with risk. Assert the outcome your system owns, not the intermediate state your infrastructure owns. When verifying async work, assert the final business state — never the delivery/retry bookkeeping held in shared infrastructure that any co-running process can write. Such a check passes when run alone and flakes the moment anything else shares that infrastructure. Keep shared guidance role-relevant. Universal guidance must help every receiving skill or agent; code-specific obligations belong only in code-specific protocols.

Nested Task Expansion Contract — For workflow-step invocation, the [Workflow] ... row is only a parent container; the child skill still creates visible phase tasks.

1. Call TaskList first. If a matching active parent workflow row exists, set nested=true and record parentTaskId; otherwise run standalone.
2. Create one task per declared phase before phase work. When nested, prefix subjects [N.M] /skill-name — phase.
3. When nested, link the parent with TaskUpdate(parentTaskId, addBlockedBy: [childIds]).
4. Orchestrators must pre-expand a child skill's phase list and link the workflow row before invoking that child skill or sub-agent.
5. Mark exactly one child in_progress before work and completed immediately after evidence is written.
6. Complete the parent only after all child tasks are completed or explicitly cancelled with reason.

Blocked until: TaskList done, child phases created, parent linked when nested, first child marked in_progress.

Project Reference Docs Gate (static JIT) — Run after task-tracking bootstrap and immediately before target/source file reads, grep, edits, tests, or analysis. Project docs override generic framework assumptions; hooks may remind or accelerate this gate, but never prove that it ran.

1. Identify scope: file types, domain area, and operation.
2. Read docs/project-config.json first — the project's machine-readable map. It is the single source of truth for THIS repo (modules/paths, framework + search keywords, test/E2E/integration run-commands, design system, architecture rules, workflow patterns); ground exact paths, run-commands, and conventions on it before investigating, planning, or coding — never assume framework defaults (CLAUDE.md + reference docs are derived from it). If it — or the docs index, lessons.md, CLAUDE.md, AGENTS.md, or any required reference doc — is missing or stale, auto-run /project-init or the narrow route (/project-config, /docs-init, /scan-all, /scan --target=<key>, /claude-md-init) first; if Codex mirrors or AGENTS.md are stale, ask the user to run /sync-codex (never auto-run it).
3. Required docs by trigger: always docs/project-reference/lessons.md; doc lookup docs-index-reference.md; review code-review-rules.md; backend/CQRS/API backend-patterns-reference.md; domain/entity domain-entities-reference.md; frontend/UI frontend-patterns-reference.md; styles/design scss-styling-guide.md + design-system/design-system-canonical.md; integration tests integration-test-reference.md; E2E e2e-test-reference.md; feature docs/specs feature-spec-reference.md + spec-system-reference.md + spec-principles.md; behavior/public-contract/spec-test-code sync workflow-spec-test-code-cycle-reference.md; derived spec index/ERD/reimplementation guides spec-system-reference.md + source Feature Specs under docs/specs/; architecture/new area project-structure-reference.md.
4. Read every required doc, then before target work state: Reference docs read: ... | Not applicable: .... After compaction, resume, delegation, or a material context change, repeat the route and restate the set; prior conversation and hook output are not proof of current loading.

Ready when: scope evaluated, docs/project-config.json consulted, required docs checked/read or setup route completed, lessons.md confirmed, citation emitted.

Task Tracking & External Report Persistence — Bootstrap this before execution; then run project-reference doc prefetch before target/source work.

1. Create a small task breakdown before target file reads, grep, edits, or analysis. On context loss, inspect the current task list first.
2. Mark one task in_progress before work and completed immediately after evidence; never batch transitions.
3. For plan/review work, create plans/reports/{skill}-{YYMMDD}-{HHmm}-{slug}.md before first finding.
4. Append findings after each file/section/decision and synthesize from the report file at the end.
5. Final output cites Full report: plans/reports/{filename}.

Blocked until: task breakdown exists, report path declared for plan/review work, first finding persisted before the next finding.

Critical Thinking Mindset — Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act. Anti-hallucination: Never present guess as fact — cite sources for every claim, admit uncertainty freely, self-check output for errors, cross-reference independently, stay skeptical of own confidence — certainty without evidence root of all hallucination.

Incremental Result Persistence — MANDATORY for all sub-agents or heavy inline steps processing >3 files.

1. Before starting: Create report file plans/reports/{skill}-{date}-{slug}.md and record Run ID, Task ID, Attempt ID, target scope, and target fingerprint.
2. After each file/section reviewed: Append findings, evidence, changed paths, and gaps immediately — never hold them in memory.
3. Delegated return: A sub-agent emits only the structured SYNC:subagent-return-contract envelope with exact totals, salient Critical/High findings (maximum ten), current attempt, and Full report: path. Inline user-facing output: Preserve the skill's requested explanation or teaching, with links to the persisted evidence; the delegated transport limit does not replace that deliverable. Do not paste a full review report into an envelope.
4. Parent synthesis: The main agent reads the full report for synthesis, acceptance, deduplication, and repair planning — not only when a named blocker exists. It preserves all severities beyond the transport cap.
5. Read-only boundary: A read-only leaf may write its report/repair proposal but MUST NOT edit source, generated output, or user data; the parent/owner performs repairs after acceptance.
6. Advancement gate: The parent records ACCEPTED for the current Attempt ID only after reconciling target, totals, gaps, and changed paths; stale or late attempts cannot advance dependent work.

Why: Context cutoff mid-execution loses ALL in-memory findings. Each disk write survives compaction. Partial results are better than no results, while explicit identity prevents a late result from being mistaken for the current run.

Report naming: plans/reports/{skill-name}-{YYMMDD}-{HHmm}-{slug}.md

Sub-Agent Return Contract — When this skill spawns a sub-agent, the sub-agent MUST return ONLY the structured envelope below. Main agent reads the envelope first, then opens the referenced report for synthesis, acceptance, deduplication, or repair planning; a full report is never pasted inline.

## Sub-Agent Result: [skill-name]

Status: ✅ PASS | ⚠️ PARTIAL | ❌ FAIL
Confidence: [0-100]%
Run ID: [stable run identifier]
Task ID: [parent task or phase identifier]
Attempt ID: [monotonic attempt/revision identifier]
Target: [exact files/paths or scope] @ [target fingerprint/commit]
Changed paths: [none | exact paths]
Finding totals: Critical=[n] | High=[n] | Medium=[n] | Low=[n]
Acceptance: PENDING | ACCEPTED | REJECTED — parent records the decision

### Findings (Critical/High surfaced — max 10 bullets)

- [severity] [file:line] [finding]

### Gaps / Unverified

- [missing host, runtime, coverage, or evidence limitation]

### Actions Taken

- [file changed] [what changed]

### Blockers (if any)

- [blocker description, or `none`]

Full report: plans/reports/[skill-name]-[date]-[slug].md

The ten-bullet limit is a transport limit, not a visibility limit: the full report may contain more than ten Medium/Low findings when no named blocker exists, and the parent MUST read it when synthesizing or deduplicating. The parent MUST reject a stale, duplicate, or superseded Attempt ID and MUST accept the current attempt before advancing a dependent step. Read-only leaves write repair proposals/reports only; they do not edit source, generated carriers, or user files.

Context budget — the return payload is a SUMMARY, not a transcript: no raw file contents / full diffs / verbatim logs inline, no re-pasted source. Everything beyond the envelope lives in the incrementally-written report. A sub-agent that would exceed the summary shape MUST persist the detail and return only the pointer; bounded transport must never become bounded visibility.

Test Architecture & Execution Contract — Treat testability as a setup/architecture acceptance condition. For every potentially applicable tier — Unit, Integration/System, E2E, and Performance/Scale (warranted at T1+/B2+) — record APPLICABLE only with evidence of its runner/framework/configuration; otherwise record N/A — <evidence> and never fabricate coverage.

1. Matrix before implementation: Record applicability, owner, runner/framework, test root, fixture/data strategy, full command, focused/partial command, zero-match behavior, CI gate, a simple/Windows entry point (a .cmd when the project needs one), the host-mode AND container-mode commands where the project supports both, and the environment reach (which of local / CI / production-shaped this tier can target).
2. Runnable scopes: Full and focused commands must be copy-ready, fail on invalid or zero-match selections, report exact counts and exit status, and be safe to repeat. E2E uses only configured browser/service commands.
3. Fresh valid state: Each run/test owns a unique run identity and business-data suffix, arranges through supported public paths, and uses realistic valid data. Reference setup is count-before-create, idempotent, and restart-safe. Intentional accumulation is additive, keyed, and integrity-checked; never hide contamination with destructive reset. Run-scoped cleanup, when supported, is opt-in and idempotent: after evidence capture it may remove only ephemeral resources owned by the current run; it must never delete persistent/additive data or another run's data, reset shared state, or replace no-reset proof.

…(truncated)