Workflow Code To Spec

Codex compatibility note:

• Invoke repository skills with $skill-name in Codex; this mirrored copy rewrites legacy Claude /skill-name references.
• Task tracker mandate: BEFORE executing any workflow or skill step, create/update task tracking for all steps and keep it synchronized as progress changes.
• User-question prompts mean to ask the user directly in Codex.
• Ignore Claude-specific mode-switch instructions when they appear.
• Strict execution contract: when a user explicitly invokes a skill, execute that skill protocol as written.
• Subagent authorization: when a skill is user-invoked or AI-detected and its protocol requires subagents, that skill activation authorizes use of the required spawn_agent subagent(s) for that task.
• Do not skip, reorder, or merge protocol steps unless the user explicitly approves the deviation first.
• For workflow skills, execute each listed child-skill step explicitly and report step-by-step evidence.
• If a required step/tool cannot run in this environment, stop and ask the user before adapting.

Codex Project-Reference Loading (No Hooks)

Codex uses static project-reference loading instead of runtime-injected project docs. When coding, planning, debugging, testing, or reviewing, open project docs explicitly using this routing.

Always read:

• docs/project-config.json (project-specific paths, commands, modules, and workflow/test settings)
• docs/project-reference/docs-index-reference.md (routes to the full docs/project-reference/* catalog)
• docs/project-reference/lessons.md (always-on guardrails and anti-patterns)

Missing/stale context route: If docs/project-config.json, the docs index, lessons.md, CLAUDE.md, AGENTS.md, or any task-required reference doc is missing or stale, auto-run $project-init or the narrow setup route ($project-config, $docs-init, $scan-all, $scan --target=<key>, $claude-md-init) before ordinary project-specific work. If Codex mirrors or AGENTS.md are missing/stale, ask the user to run $sync-codex; do not auto-run it.

Situation-based docs:

• Project structure/architecture/tech-stack/deployment/setup (any layer — backend, frontend, or infra): project-structure-reference.md
• Backend/CQRS/API/domain/entity changes: backend-patterns-reference.md, domain-entities-reference.md
• Frontend/UI/styling/design-system: frontend-patterns-reference.md, scss-styling-guide.md, design-system/README.md
• Spec authoring, docs/specs/ pathing, or TC format: feature-spec-reference.md, spec-system-reference.md, spec-principles.md
• Behavior/public-contract changes or spec-test-code sync: workflow-spec-test-code-cycle-reference.md plus the spec docs above
• Derived spec indexes/ERDs/reimplementation guides: spec-system-reference.md and source Feature Specs under docs/specs/
• Integration test implementation/review: integration-test-reference.md
• E2E test implementation/review: e2e-test-reference.md
• Code review/audit work: code-review-rules.md plus domain docs above based on changed files

Do not read all docs blindly. Start from docs-index-reference.md, then open only relevant files for the task.

[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.

Renamed: formerly workflow-build-specs — now $workflow-code-to-spec. The old name no longer resolves as a slash command.

Quick Summary

Goal: Keep the Feature Spec, implementation, tests, and project docs synchronized through a governed spec-driven workflow — a single entry point for spec-driven doc generation and maintenance over ONE canonical artifact, the tech-free 8-section Feature Spec at docs/specs/{Bucket}/README.{Feature}.md (code is the technical source of truth; there is no parallel engineering tree).

[SINGLE HOME] There is ONE canonical artifact — the tech-free 8-section Feature Spec authored by spec at docs/specs/{Bucket}/. There is no parallel A-E "Engineering Spec" bundle and no separate Business Feature Docs tree; spec-index only regenerates a DERIVED index/ERD over the Feature Specs. Authority: docs/project-reference/spec-system-reference.md.

One Canonical Artifact + Derived Aids

App Bucket Mapping

Resolve service→bucket assignments from the canonical table in docs/project-reference/spec-system-reference.md → App Bucket Mapping — do not inline project-specific bucket names in this skill.

Mode Routing:

Key Rules:

• MUST ATTENTION define success criteria before execution and loop until observable verification passes.
• MUST ATTENTION when creating/reviewing specs or tests, name Business Intent / Invariant Guarded or the protected business intent/invariant and ensure the test would fail if that intent breaks.
• Confirm mode by asking the user directly BEFORE any action — NEVER skip Step 0
• Invoke skill invocation for EACH step — NEVER batch-complete or mark done without invocation
• Spawn sub-agents for 4+ capabilities in ONE message — NEVER sequential
• Every spec carries the THREE layers needed for stack-agnostic rebuild: business logic (§1-7), UI/UX interaction surface (§6.2-6.5 for UI-bearing features), and test specs (§8) — a spec missing any layer is incomplete
• Trace the FULL vertical chain (UI → API → handler → domain → event → read model → UI) per capability and reconcile it (Step 1-INIT.4.6) — a layer-parallel inventory silently misses seam behavior; nothing-missed requires the end-to-end trace
• For multi-bucket/whole-project scope, maintain the resumable Coverage Ledger (Step B.3) and clear the Whole-Project Completeness Gate before $workflow-end
• §1-7 of every Feature Spec are STRICTLY tech-free; the §5 Mermaid ERD is authored inside the Feature Spec — there is no separate ERD file
• Write findings incrementally after each section — NEVER hold in memory
• If shared skills/workflows/hooks/sync tooling changed, run npm run codex:sync before $workflow-end or record explicit N/A evidence; verify generated mirrors are current.

Step 0 — Mode Detection (MANDATORY FIRST)

Use ask the user directly to confirm mode before any action.

Auto-detection rules:

IF docs/specs/{Bucket}/ has NO Feature Spec for the target scope
  → Suggest: init-full

IF git diff has service/frontend changes touching an already-spec'd capability
  → Suggest: update

IF explicit --audit flag OR user says "audit" / "check freshness" / "are docs stale"
  → Suggest: audit

Bucket + capability confirmation:

• Probe Glob docs/specs/{Bucket}/README.*.md to see which capabilities already have a Feature Spec.
• Map the changed services to a Bucket via the App Bucket Mapping table above.
• Confirm the capability name (PascalCase) with the user — this becomes README.{Feature}.md.

Present the detected mode with reasoning. User confirms before proceeding.

MODE: init-full

When to Use

Starting from zero: no docs/specs/{Bucket}/README.{Feature}.md for the target scope.

Step Sequence

## Step A — Discovery (scout)

$scout
  → Holistic codebase map — capability registry, entry points, integration boundaries
  → Identify the set of capabilities in scope (one Feature Spec each)
  → **Full-chain awareness:** map not just backend layers but the complete vertical slice per
     capability — UI view/action → API → handler → domain entity/rule → event → consumer/read model
     → UI outcome. The detailed per-capability trace + reconciliation gate runs inside
     `$spec [mode=init]` Step 1-INIT.4.6 (use `graph-connect-api` for frontend→backend links,
     `graph-trace` for backend flow); scout just confirms both ends (frontend + backend) are in scope
     so no seam is invisible at planning time.
  → Task tracking: "scout — enumerate capabilities + full FE→BE→domain entry points for {Bucket}"

## Step B — Size Evaluation and Divide-and-Conquer Planning (MANDATORY — runs BEFORE plan)

**[BLOCKING GATE]** Before writing any plan, evaluate scope and recursively decompose until
each capability is an independently authorable Feature Spec. **The task may be huge — an entire
project's worth of code-to-spec is the explicit worst case. Size it honestly first, then break it
into many small, independently-completable units; never start authoring against an unsized scope.**

### B.1 — Classify scope breadth FIRST

| Scope | Signal | Strategy |
| ----- | ------ | -------- |
| **Single capability** | one feature named / one diff | single-session authoring |
| **Single bucket** | one module/service | per-capability decomposition within the bucket |
| **Whole project** | "all specs", "the whole project", "every feature", no bucket named | **enumerate ALL buckets × ALL capabilities → build the Coverage Ledger (B.3) → run bucket-by-bucket across resumable sessions** |

### B.2 — Estimate per run

```
capability_count = count of distinct capabilities in the current scope
  IF capability_count > 10:  → SPLIT into groups: max 10 capabilities per run, run groups sequentially
  IF 4 ≤ capability_count ≤ 10: → Sub-agents mandatory (one spec sub-agent per capability)
  IF capability_count ≤ 3:   → Single-session authoring
```

**Recursive decomposition rule:** Feature Specs have no line-count cap. If a single capability has
>40 TCs or contains distinct module-level capabilities, split the capability into sibling specs.

### B.3 — Whole-Project Coverage Ledger [BLOCKING for whole-project scope; recommended for multi-bucket]

When the scope spans more than one bucket, a single session cannot hold it — so coverage MUST be
tracked in a **persistent, resumable ledger on disk**, not in memory or the task list alone. This is
the guard for "the whole-project code-to-spec works flawlessly and nothing is missed."

1. **Enumerate the full target set:** for every bucket (App Bucket Mapping), list every capability
   discovered in scout. This is the denominator — the complete set that MUST end with a Feature Spec.
2. **Write the ledger** to `plans/reports/code-to-spec-coverage-{date}.md`:

   | Bucket | Capability | Spec path | §1-7 | §6 (UI) | §8 TCs | Chain trace | Status |
   | ------ | ---------- | --------- | :--: | :-----: | :----: | :---------: | ------ |
   | {Bucket} | {Capability} | docs/specs/{Bucket}/README.{Feature}.md | ⬜ | ⬜ | 0 | ⬜ | NOT STARTED |

3. **Update the ledger row after each capability completes** (incremental persistence — never batch).
   `Status` ∈ NOT STARTED → IN PROGRESS → SPEC DONE → REVIEWED. `§6 (UI)` is `n/a` for backend-only.
4. **Resume rule:** on any session start / after compaction, read the ledger FIRST, re-glob
   `docs/specs/**` to confirm, and continue from the first non-REVIEWED row — NEVER re-author a done
   capability, NEVER re-run scout for already-enumerated buckets.
5. **[BLOCKING] Whole-Project Completeness Gate (before `$workflow-end`):** every row is `REVIEWED`,
   OR carries an explicit, recorded deferral reason. A bucket/capability discovered in scout but
   absent from `docs/specs/**` with no deferral reason = the workflow is NOT complete. Report the
   final coverage as `{reviewed}/{total} capabilities` in `$watzup`.

Task tracking: "size-evaluation — classify scope breadth, count capabilities, build coverage ledger (whole-project), decide split strategy"

## Step C — Plan

$plan
  → ONE task per capability Feature Spec (author §1-7 + §8 placeholder)
  → Core domain capabilities first, supporting ones last
  → Verify the current task list count ≥ capability_count before proceeding (BLOCKING gate)
  → Task tracking: "plan — produce per-capability authoring plan"

## Step D — Plan Review

$plan-review
  → Validate: every in-scope capability has an authoring task, no gaps
  → Verify: no sub-agent handles >3 capabilities; each capability within caps
  → Task tracking: "plan-review — validate capability coverage + caps"

## Step E — Plan Validation

$plan-validate
  → User confirms: bucket, capability list, scope, split strategy
  → Task tracking: "plan-validate — user confirms scope and split strategy"

$spec [mode=init]
  → Author the canonical tech-free 8-section Feature Spec PER capability:
      §1 Overview · §2 Glossary · §3 User Stories & AC · §4 Business Rules ·
      §5 Domain Model (Mermaid ERD — MANDATORY, authored INSIDE this file) ·
      §6 Process Flows · §7 Permissions & Roles · §8 Test Specifications
  → §1-7 STRICTLY tech-free; identifiers live only in §8 evidence carriers + `[Source: ns/service/id]` + ` ```mermaid ``` ` blocks
  → Output: docs/specs/{Bucket}/README.{Feature}.md + bucket INDEX.md
  → Sub-agents for 4+ capabilities (BLOCKING: ONE message spawn); each prompt includes capability name, output path, tech-agnostic contract, SYNC protocols
  → No line-count cap applies — split when TCs>40 or distinct module-level capabilities emerge

$spec [mode=tests]
  → Populate Section 8 with demoable business TC-{FEATURE}-{NNN} entries (BDD, Business Intent, abstract Evidence, CoveredBy field, Status)

$artifact-review --type=spec-tests
  → Review §8 TCs for invariant coverage, GIVEN/WHEN/THEN completeness, duplicate TC IDs

$artifact-review
  → Quality check the Feature Spec(s):
    - §1-7 strictly tech-free (zero framework/product/language terms in prose)
    - §5 Mermaid ERD present with entities, relationships, cardinalities
    - §8 every TC has Business Intent, abstract `[Source: ns/service/id]` anchor, CoveredBy field, Status
    - YAML frontmatter present; no line-count cap applied
  → PASS criteria: zero [UNVERIFIED] without exclusion reason + zero tech terms in §1-7
  → Gap found → validate findings → fix validated gaps → restart full artifact-review pass from the first check

$docs-update
  → Near-final synchronization sweep across project docs, the Feature Spec(s), and Section 8
  → MUST run after artifact-review fixes and before $workflow-end
  → (Optional) regenerate the derived bucket INDEX / ERD via $spec-index
  → Report skipped sub-phases explicitly when no impacted docs exist

$workflow-end

$watzup
  → Session summary: capabilities authored, files written, ~lines, §8 TC counts, coverage gaps, open questions (confidence < 80%), plus final $understand handoff

MODE: update

When to Use

Code changed (new feature, bug fix, refactor, new PBI). Sync the Feature Spec incrementally.

Scope Sources

1. Auto-detect from git diff --name-only HEAD (default)
2. Explicit capability list from user
3. New PBI or requirement description (map to affected capabilities manually)

Step Sequence

$workflow-review-changes
  → Full code review cycle + docs-update (Phase 2 spec diff-scoped sync)
  → Produces: impact map (capabilities affected, Feature Spec sections to update)

$spec [mode=update]
  → Update only the impacted sections of each affected Feature Spec (full 8-section pass with 3-pass verification when restructuring)
  → SKIP if docs-update Phase 2 completed with zero gaps — mark "Skipped: docs-update Phase 2 sufficient"

$spec [mode=tests]
  → **EXPLICIT TC STEP — required when new functionality added**
  → SKIP if changes are purely cosmetic (CSS, comments, config) — mark "Skipped: no behavioral impact"
  → Mode detection: new commands/queries/endpoints → implement-first; PBI-driven → TDD-first; TC edits only → update
  → Write/update TCs in Section 8; grep existing IDs before assigning; never overwrite Tested TCs

$artifact-review --type=spec-tests
  → Review updated/planned TCs for invariant coverage, GIVEN/WHEN/THEN completeness, stale TC handling, duplicate IDs
  → SKIP if $spec [mode=tests] skipped — mark "Skipped: no TC changes"
  → BLOCK sync if review finds missing invariants, ambiguous behavior, or TC/code/spec drift

$spec [mode=sync]
  → Refresh the derived bucket INDEX TC counts from Section 8 (Section 8 stays canonical — no separate dashboard)
  → SKIP if no TC changes in this cycle

> **UI-intent maintenance (conditional)** — runs alongside the `$spec [mode=sync]` / spec authoring step, **only when the change carries user-facing behavior** (else state the skip reason — backend-only change, no §6 change). When the reverse-engineered code carries user-facing behavior, the spec authoring step MUST refresh the Feature Spec **§6** interaction surface from the implemented capability — View Inventory, Key UI States, and the per-story (`US-`/`OP-`/`BR-`) click-path — and link the governing `$design-spec`/mockup in the spec frontmatter so §6 and the design artifact stay coupled to what the code actually does. The rules live in the shared block below (`SYNC:ui-intent-layer`) — follow it; do not restate it here.

$changes-review
  → Holistic review of Feature Spec changes
  → Verify: spec changes match code changes (no over/under documentation); §8 covers all new functionality

$docs-update
  → Near-final synchronization sweep across project docs, the Feature Spec(s), and Section 8
  → MUST run after changes-review fixes and before $workflow-end
  → Report skipped sub-phases explicitly when no impacted docs exist

$workflow-end

$watzup
  → Summary: capabilities updated, Feature Spec sections changed, TCs added/updated, new open questions, plus final $understand handoff

New PBI / Requirement Update Protocol

Triggering update from new PBI or requirement (not code change):

After $workflow-review-changes:
  → User provides PBI text or requirement description
  → Map requirement → affected domain entities → affected capabilities
  → $dor-gate: required when a new/changed PBI is being made implementation-ready; PASS or WARN before planned specs/TCs become guidance
  → $pbi-mockup: required when the PBI changes user-facing UI/journeys; skip with reason for backend-only requirements
  → Treat as "speculative update": add new rules/sections to the Feature Spec marked [PLANNED — not yet implemented]
  → Add corresponding TCs in Section 8 marked Status: Planned
  → $artifact-review --type=spec-tests: review planned TCs before refreshing the derived index
  → These become implementation guidance, not verified spec

MODE: audit

When to Use

Periodic health check (quarterly or before major release). Verify the Feature Specs are current.

Audit Time Estimation

audit_effort = capability_count × 8min_per_capability
example: 8 capabilities × 8min = ~1h (AI-assisted)

Budget multiplier: If last audit was >90 days ago → ×1.5 (more drift expected).

Step Sequence

$scout
  → Quick codebase scan: current state of entities, commands, controllers (lightweight, 30min max)

$spec [mode=audit]
  → Compare each Feature Spec's last_updated vs git log of the source it documents
  → Output: stale capabilities/sections table with recommended update scope

$artifact-review
  → Consolidated audit report
  → Produce: plans/reports/spec-audit-{date}-{Bucket}.md
  → Include: total stale coverage %, estimated update effort, priority order
  → (Optional) $spec-index [mode=audit] — report which DERIVED index/ERD aids lag their source specs

$docs-update
  → Near-final synchronization sweep; MUST run after artifact-review conclusions and before $workflow-end
  → Report skipped sub-phases explicitly when no impacted docs exist

$workflow-end

$watzup
  → Present action plan: which capabilities to update first
  → Recommend: run update mode scoped to stale capabilities
  → Run final $understand handoff

Conditional Skip Rules

Sub-Agent Coordination Protocol (init-full, 4+ capabilities)

1. $scout + $plan in main context → capability registry + per-capability task list
2. Spawn spec sub-agents (one per capability) in ONE message
3. Wait for all sub-agents to complete
4. Spawn spec [mode=tests] sub-agents (one per capability) in ONE message to populate Section 8
5. Main context assembles + verifies in $artifact-review

Each spec sub-agent receives:

• Capability name + bucket
• Output path: docs/specs/{Bucket}/README.{Feature}.md
• Tech-agnostic contract (§1-7 tech-free; §5 ERD mandatory)
• Incremental persistence instruction (write after each section)

Integration with docs-update workflow step

docs-update called as a workflow step (not standalone) synchronizes the single canonical artifact:

docs-update (as workflow step):
  Phase 1: Project docs (inline — unchanged)
  Phase 2: $spec update mode (impacted sections of the Feature Spec)
  Phase 3: $spec [mode=tests] (Section 8 TCs)
  Phase 3.5: $artifact-review --type=spec-tests (required when Phase 3 changes TCs)
  Phase 4: $spec [mode=sync] (refresh derived INDEX TC counts)
  Phase 4.5: $spec-index (OPTIONAL — regenerate derived bucket INDEX / ERD if one is maintained)

The Feature Spec stays in sync on every feature/bugfix/refactor workflow.

When to Use vs When NOT to Use

Use This Workflow

• First-time Feature Spec authoring for a capability or full bucket
• After significant code changes (new feature, major refactor)
• Onboarding new team to a capability — the Feature Spec as knowledge handoff artifact
• Before tech migration or re-implementation (pair with a derived reimplementation guide)
• Quarterly spec health audits
• After new PBIs groomed for implementation
• Compliance documentation — prove system behavior in plain language
• Verify design intent — compare the Feature Spec against original vision

Use Standalone Skills Instead

[TASK-PLANNING] Before acting, analyze task scope and systematically break it into small todo tasks and sub-tasks using task tracking.

IMPORTANT MANDATORY Steps: $scout -> $plan -> $plan-review -> $plan-validate -> $spec -> $spec [mode=tests] -> $artifact-review --type=spec-tests -> $artifact-review -> $docs-update -> $workflow-end -> $watzup

[BLOCKING] Invoke skill invocation for EACH step — NEVER batch-complete, NEVER mark done without skill invocation. [BLOCKING] Confirm mode by asking the user directly BEFORE any action — NEVER skip Step 0. [BLOCKING] Spawn sub-agents for 4+ capabilities in ONE message — NEVER sequential spawning. [BLOCKING — Context Compaction / Session Resume] At any session start or after context compaction: (1) the current task list FIRST — resume existing, NEVER create duplicates; (2) re-glob docs/specs/{Bucket}/ to see which capabilities already have a Feature Spec — skip those; (3) NEVER re-run $scout or $plan in a resumed session. [BLOCKING] Read docs/project-reference/spec-principles.md before running any author/update/audit step — it is the shared spec quality baseline (tech-agnostic rule + banned-token list).

Goal Contract propagation (workflow-owned): At workflow start, 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 spec request). Map each spec/test/code cycle output (Feature Specs authored, TCs written, audit findings fixed) to the saved success criteria and append the evidence to the goal file's Iteration Log per cycle. Before $workflow-end, emit the Goal Satisfaction matrix (PASS/FAIL/BLOCKED); completion requires every required criterion PASS or BLOCKED with a user-facing escalation.

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 the current task list 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: the current task list done, child phases created, parent linked when nested, first child marked in_progress.

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.

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.

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
2. After each file/section reviewed: Append findings to report immediately — never hold in memory
3. Return to main agent: Summary only (per SYNC:subagent-return-contract) with Full report: path
4. Main agent: Reads report file only when resolving specific blockers

Why: Context cutoff mid-execution loses ALL in-memory findings. Each disk write survives compaction. Partial results are better than no results.

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 this structure. Main agent reads only this summary — NEVER requests full sub-agent output inline.

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

Status: ✅ PASS | ⚠️ PARTIAL | ❌ FAIL
Confidence: [0-100]%

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

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

### Actions Taken

- [file changed] [what changed]

### Blockers (if any)

- [blocker description]

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

Main agent reads Full report file ONLY when: (a) resolving a specific blocker, or (b) building a fix plan. Sub-agent writes full report incrementally (per SYNC:incremental-persistence) — not held in memory.

Context budget — the return payload is a SUMMARY, not a transcript: ≤10 finding bullets, no raw file contents / full diffs / verbatim logs inline, no re-pasted source. Everything beyond the summary lives in the Full report on disk. A sub-agent that would exceed the summary shape MUST write the detail to its report and return only the pointer — the orchestrator's context is the scarce resource the whole map-reduce protects.

[BLOCKING] Capture a tech-agnostic UI/UX intent layer in every UI-bearing spec — a reader must be able to visualize how the feature works without naming any technology. When the feature has a user interface, the spec MUST ATTENTION carry an interaction-surface section so the application — not just its API — can be rebuilt on any stack:

1. View Inventory — list each view/screen by its UX ROLE and purpose (e.g. "list of items", "item editor", "confirmation step") and what information it presents. Describe by role, never by an implementation name.
2. Navigation Map — how a user moves between views: entry points, transitions, and exits. Trace how this surface connects to neighboring features already in the system.
3. Key observable UI States — the distinct states a user can observe per view (empty, loading, populated, error, success, permission-denied, etc.) — described as what the user perceives, not how it is rendered.
4. Per-story interaction flow — for each user story, the step-by-step click/action path from intent to outcome, cross-referenced to the logical IDs the spec already owns (US-/OP-/BR-).
5. Couple to the companion design artifact — keep deep visual fidelity (layout, tokens, pixel detail) OUT of the spec; it lives in the linked design-spec/mockup. Record that companion's path in the spec frontmatter so the spec stays the navigable hub.

M1-clean (NON-NEGOTIABLE): the prose names ZERO frameworks, routes/URLs, CSS, or component-class names — only roles, information, states, and flows. Technology detail belongs in the companion design artifact, never here.

Skip ONLY when the feature is backend-only (no UI) — state that reason explicitly in the section.

• MANDATORY Parent workflow rows do not replace child phase tracking; expand phases and link the parent when nested.
• MANDATORY Orchestrators pre-expand child skill phases before invocation; use [N.M] $skill-name — phase prefixes and one-in_progress discipline.

• MANDATORY Resolve the active Goal Contract BEFORE work (active plan goal.md → plans/goals/{YYMMDD-HHmm}-{slug}/goal.md → create from current request) and read saved success criteria before editing.
• MANDATORY Append iteration evidence after execution; emit a Goal Satisfaction matrix (PASS/FAIL/BLOCKED) before reporting PASS; loop on validated FAIL; escalate repeated no-progress or blockers. NEVER store secrets in goal files.

• MANDATORY For UI-bearing specs, author/maintain the tech-agnostic interaction-surface layer (View Inventory + Navigation Map + observable UI States + per-story US-/OP-/BR--traced flow); keep deep visual fidelity in the linked design-spec/mockup recorded in frontmatter; name ZERO frameworks/routes/CSS/component classes; skip ONLY for backend-only features with a stated reason.

Prompt-Enhance Closing Anchors

IMPORTANT MUST ATTENTION follow declared step order for this skill; NEVER skip, reorder, or merge steps without explicit user approval IMPORTANT MUST ATTENTION for every step/sub-skill call: set in_progress before execution, set completed after execution IMPORTANT MUST ATTENTION every skipped step MUST include explicit reason; every completed step MUST include concise evidence IMPORTANT MUST ATTENTION if Task tools unavailable, maintain an equivalent step-by-step plan tracker with synchronized statuses

Parallel Sub-Agent Dispatch — Plan parallelism the moment a task breakdown exists, BEFORE executing it — running provably independent tasks sequentially wastes wall-clock. Applies to every multi-step job: workflow steps, planning, batch updates, investigation, research, scans, reviews, doc sync. Plan execution is metadata-gated, NEVER default-parallel — fan-out follows ONLY what the plan declares (PAR/SEQ tags + per-phase write set); an untagged plan runs sequentially — why: a derived write set cannot see cascade or generated writes.

1. Tag every task PAR or SEQ. PAR = inputs exclude every pending task's output AND write set disjoint from every other PAR. Else SEQ — MUST ATTENTION name the dependency forcing it.
2. Group PAR into waves. No edge between members. Two writers of one file NEVER share a wave. Read-only work (search, investigation, review, research) parallelizes freely.
3. Declare before dispatch: Parallel plan: wave 1 = [...] · wave 2 = [...] · SEQ = [...] (reason).
4. Spawn each wave in ONE message — every spawn_agent call in one response, NEVER dripped per turn. Route each task to its specialist (.claude/skills/shared/sub-agent-selection-guide.md); NEVER code-reviewer as catch-all.
5. Brief each sub-agent self-contained: goal · scope + owned files · reference docs · return contract (summary + Full report: path, per SYNC:subagent-return-contract) · incremental persistence to plans/reports/ (per SYNC:incremental-persistence).
6. Barrier per wave. Advance ONLY after EVERY member returns (a skipped conditional counts as returned). Merge, mark each task completed/skipped, THEN dispatch the next wave. Mut

…(truncated)