claude-code-arcane 2.9.7 → 2.9.9

package/rules/{ai-code.md → llm-code.md}

@@ -1,9 +1,9 @@
 ---
 paths:
-  - "src/ai/**"
+  - "src/**/ai/**"
+  - "src/**/llm/**"
+  - "src/**/agents/**"
   - "src/ml/**"
-  - "src/llm/**"
-  - "src/agents/**"
   - "src/inference/**"
   - "prompts/**"
   - "models/**"
@@ -12,7 +12,7 @@ paths:
 # AI / LLM / ML Code Rules
 
 Applies to software integrations with LLMs, ML inference, and AI agents — not NPC behavior in games
-(that rule lives in `skills-gamedev/_rules/ai-code.md`).
+(the gamedev profiles ship their own `ai-code` rule for that).
 
 ## Prompts & Templates
 
@@ -28,7 +28,7 @@ Applies to software integrations with LLMs, ML inference, and AI agents — not
 - **Cost budget per request is measured and logged**: token counts in + out, model, cost in $. Track P50/P95 cost per feature.
 - **Timeouts and retries are bounded**: max retries ≤ 2, exponential backoff, circuit breaker on the provider. A hung LLM call must not hang a user request.
 - **Streaming for user-facing generation**; batch / async for back-office work.
-- **Cache deterministic calls** (same system + same user + temperature 0 → cache the response). Document TTL and invalidation.
+- **Cache deterministic calls** (same model + same system + same user → cache the response). Document TTL and invalidation.
 
 ## Safety & PII
 

package/skills/agent-designer/README.md

@@ -377,7 +377,8 @@ with open('my_tools_anthropic.json') as f:
 # Use with Anthropic tool use
 client = anthropic.Anthropic()
 response = client.messages.create(
-    model="claude-3-opus-20240229",
+    model="claude-opus-5-5",
+    max_tokens=16000,
     messages=[{"role": "user", "content": "Search for AI news"}],
     tools=schemas['tools']
 )

package/skills/agent-designer/agent_evaluator.py

@@ -220,11 +220,6 @@ class AgentEvaluator:
     def _define_cost_benchmarks(self) -> Dict[str, Any]:
         """Define cost benchmarks for different operations"""
         return {
-            "token_costs": {
-                "gpt-4": {"input": 0.00003, "output": 0.00006},
-                "gpt-3.5-turbo": {"input": 0.000002, "output": 0.000002},
-                "claude-3": {"input": 0.000015, "output": 0.000075}
-            },
             "operation_costs": {
                 "simple_task": 0.005,
                 "complex_task": 0.050,

package/skills/ai-sdk-setup/SKILL.md

@@ -15,7 +15,7 @@ allowed-tools: Read, Glob, Grep, Bash, Write, Edit, Task
 ### Step 0: Gather Requirements
 1. **Caso:** chat / generación one-shot / extracción structured / agente con tools / RAG
 2. **Streaming** al cliente o respuesta completa
-3. **Modelo:** Opus 4.8 (`claude-opus-4-8`, razonamiento) / Sonnet 4.6 (`claude-sonnet-4-6`, balance) / Haiku 4.5 (`claude-haiku-4-5-20251001`, rápido/barato)
+3. **Modelo:** Opus 5.5 (`claude-opus-5-5`, razonamiento) / Sonnet 5.5 (`claude-sonnet-5-5`, balance) / Haiku 5.5 (`claude-haiku-5-5`, rápido/barato)
 4. ¿Necesita tool calling / structured output / RAG (ver skill `pgvector-search`)?
 
 ### Step 1: Instalar
@@ -33,7 +33,7 @@ import { streamText, convertToModelMessages } from 'ai';
 export async function POST(req: Request) {
   const { messages } = await req.json();
   const result = streamText({
-    model: anthropic('claude-sonnet-4-6'),
+    model: anthropic('claude-sonnet-5-5'),
     messages: convertToModelMessages(messages),
   });
   return result.toUIMessageStreamResponse();
@@ -44,7 +44,7 @@ Cliente: `useChat()` de `@ai-sdk/react`.
 ### Step 3: Structured output (Zod)
 ```ts
 const { object } = await generateObject({
-  model: anthropic('claude-sonnet-4-6'),
+  model: anthropic('claude-sonnet-5-5'),
   schema: z.object({ sentiment: z.enum(['pos','neg','neutral']), score: z.number() }),
   prompt: text,
 });
@@ -53,7 +53,7 @@ const { object } = await generateObject({
 ### Step 4: Tool calling / agente
 ```ts
 const result = await generateText({
-  model: anthropic('claude-opus-4-8'),
+  model: anthropic('claude-opus-5-5'),
   tools: {
     search: tool({
       description: 'Buscar docs', inputSchema: z.object({ q: z.string() }),

package/skills/architecture-decision/SKILL.md

@@ -18,11 +18,10 @@ allowed-tools: Read, Glob, Grep, Write, Task, AskUserQuestion
 
 ## Phase 0: Load Engine Context (ALWAYS FIRST)
 
-1. Read `docs/engine-reference/[engine]/VERSION.md` → engine name, version, cutoff risk
+1. Read the engine version from the project: `ProjectSettings/ProjectVersion.txt` (Unity) or `EngineAssociation` in the `.uproject` (Unreal). For Unreal, also read `docs/unreal/project-context.md` if `/ue-project-context` has run
 2. Identify domain from title (Physics, Rendering, UI, Audio, etc.)
-3. Read module reference + `breaking-changes.md` + `deprecated-apis.md`
-4. Display knowledge gap warning if MEDIUM/HIGH risk
-5. No engine configured → prompt to run `/setup-engine`
+3. If the engine version is newer than what you know reliably, say so before proposing engine APIs
+4. No engine detected → ask the user which engine and version
 
 ## Phase 1: Context Gathering
 

package/skills/architecture-review/SKILL.md

@@ -25,7 +25,7 @@ model: opus
 ## Phase 1: Load
 
 1. **L0 Summary Scan**: Grep `## Summary` from all GDDs and ADRs (fast, low tokens)
-2. **L1/L2 Full Load** (per mode): GDDs, ADRs, engine reference docs, `technical-preferences.md`, `consistency-failures.md`
+2. **L1/L2 Full Load** (per mode): GDDs, ADRs, engine and version from `ProjectSettings/ProjectVersion.txt` (Unity) or the `.uproject` (Unreal), `consistency-failures.md`
 3. Report: "Loaded [N] GDDs, [M] ADRs, engine: [name+version]"
 
 ## Phase 2: Extract Technical Requirements

package/skills/asset-audit/SKILL.md

@@ -92,4 +92,3 @@ This skill is read-only — it produces a report but does not write files.
 
 - Fix naming violations using the patterns defined in CLAUDE.md.
 - Delete confirmed orphaned assets after manual review.
-- Run `/content-audit` to cross-check asset counts against GDD-specified requirements.

package/skills/balance-check/SKILL.md

@@ -111,7 +111,7 @@ If yes:
 - Guide the user to update the relevant data file in `assets/data/` or formula in `design/balance/`
 - After each fix, offer to re-run the relevant balance checks to verify no new outliers were introduced
 - If the fix changes a tuning knob defined in a GDD or referenced by an ADR, remind the user:
-  > "This value is defined in a design document. Run `/propagate-design-change [path]` on the affected GDD to find downstream impacts before committing."
+  > "This value is defined in a design document. Run `/consistency-check` after updating the GDD to find the documents that depend on it before committing."
 
 If no:
 - Summarize open issues and suggest saving the report to `design/balance/balance-check-[system]-[date].md` for later

package/skills/brainstorm/SKILL.md

@@ -11,7 +11,7 @@ allowed-tools: Read, Glob, Grep, Write, WebSearch, Task, AskUserQuestion
 
 ## Setup
 
-1. Parse argument for genre/theme hint (`roguelike`, `space survival`, `open`). Resolve review mode: `--review` flag → `production/review-mode.txt` → default `lean`. See `.claude/docs/director-gates.md`.
+1. Parse argument for genre/theme hint (`roguelike`, `space survival`, `open`). Resolve review mode: `--review` flag → `production/review-mode.txt` → default `lean`.
 2. Check existing work: `design/gdd/game-concept.md` (resume), `design/gdd/game-pillars.md` (build on).
 3. Run phases **interactively** — collaborative exploration, not silent generation. Use `AskUserQuestion` at key decision points.
 
@@ -74,12 +74,12 @@ Using Bartle taxonomy + Quantic Foundry: primary player type, secondary appeal,
 ### Phase 6: Scope and Feasibility
 
 - **Platform**: AskUserQuestion (PC / Mobile / Console / Web / Multiple)
-- **Engine**: AskUserQuestion (Unity / Unreal / No preference → suggest `/setup-engine` later). Don't second-guess if they pick one.
+- **Engine**: AskUserQuestion (Unity / Unreal / No preference → once decided, `/scaffold-unity` for Unity or `/ue-project-context` for Unreal). Don't second-guess if they pick one.
 - **Art pipeline, content scope, MVP definition, biggest risks, scope tiers**
 
 ## Document Generation
 
-Generate game concept using template at `.claude/docs/templates/game-concept.md`. Include Visual Identity Anchor section. Scope consistency rule: Estimated Scope must match full-vision timeline.
+Generate game concept using template at `.claude/skills/_templates/game-concept.md`. Include Visual Identity Anchor section. Scope consistency rule: Estimated Scope must match full-vision timeline.
 
 AskUserQuestion for write approval → if "revise" → iterate until approved → write to `design/gdd/game-concept.md`.
 

package/skills/brainstorm/references/next-steps-pipeline.md

@@ -1,13 +1,12 @@
 # Next Steps — Full Pipeline
 
-1. `/setup-engine` — configure engine + reference docs
+1. `/scaffold-unity` (Unity) or `/ue-project-context` (Unreal) — set up the engine project
 2. `/art-bible` — visual identity (BEFORE GDDs)
-3. `/design-review design/gdd/game-concept.md` — validate completeness
-4. `/map-systems` — decompose into systems with dependencies
-5. `/design-system` — per-system GDDs
-6. `/create-architecture` — master architecture blueprint + ADR list
-7. `/architecture-decision` (xN) — one ADR per decision
-8. `/gate-check` — phase gate before production
-9. `/prototype [core-mechanic]` — validate core loop
-10. `/playtest-report` — validate hypothesis
+3. `/map-systems` — decompose into systems with dependencies
+4. `/design-system` — per-system GDDs
+5. `/consistency-check` — cross-GDD conflicts before architecture
+6. `/architecture-decision` (xN) — one ADR per decision
+7. `/gate-check` — phase gate before production
+8. `/prototype [core-mechanic]` — validate core loop
+9. `/playtest-report` — validate hypothesis
 11. `/sprint-plan new` — first sprint

package/skills/consistency-check/SKILL.md

@@ -10,7 +10,7 @@ allowed-tools: Read, Glob, Grep, Write, Edit, Bash
 
 Detects cross-document inconsistencies by comparing GDDs against entity registry (`design/registry/entities.yaml`). Grep-first: read registry once, then grep only matching names — no full document reads.
 
-Run: after each new GDD, before `/review-all-gdds`, before `/create-architecture`. Output: conflict report + optional registry corrections.
+Run: after each new GDD, and before `/architecture-decision`. Output: conflict report + optional registry corrections.
 
 ## Arguments
 
@@ -62,7 +62,7 @@ If 🔴 conflicts found → append to `docs/consistency-failures.md` (if exists)
 
 ## Protocol
 
-- PASS → proceed to `/review-all-gdds` or `/create-architecture`
+- PASS → proceed to `/architecture-decision`
 - CONFLICTS FOUND → fix GDDs, re-run to confirm
 - Source GDD (per `source:` field) is authoritative owner
 - Run after each new GDD to catch issues early

package/skills/design-system/SKILL.md

@@ -23,7 +23,7 @@ Read all before asking anything:
 4. `design/registry/entities.yaml` — known cross-system facts (locked values)
 5. `docs/consistency-failures.md` — past conflict patterns for this domain
 6. `design/gdd/game-pillars.md`, related GDDs (optional)
-7. `.claude/docs/technical-preferences.md` → engine + version → engine reference docs → feasibility brief
+7. Engine and version from `ProjectSettings/ProjectVersion.txt` (Unity) or the `.uproject` (Unreal) → feasibility brief
 
 Present context summary with dependencies, registry facts, engine constraints. Warn if upstream deps undesigned.
 

package/skills/design-system/references/post-design-validation.md

@@ -3,6 +3,5 @@
 1. **Self-check**: All 8 sections have real content, formulas have variables, edge cases have resolutions
 2. **CD Pillar Review** (full mode only): Spawn `creative-director` gate CD-GDD-ALIGN
 3. **Registry update**: Scan for new entities/items/formulas/constants -> append to `design/registry/entities.yaml`
-4. **Offer review**: Direct to `/design-review` in **fresh session** (never inline)
-5. **Update systems index**: Status -> Designed/Approved/In Review
-6. **Suggest next**: `/consistency-check`, next system, `/gate-check`
+4. **Update systems index**: Status -> Designed/Approved/In Review
+5. **Suggest next**: `/consistency-check`, next system, `/gate-check`

package/skills/frontend-dev/SKILL.md

@@ -159,7 +159,7 @@ project/
 2. Plan motion sequences following performance guardrails
 
 ### Phase 3: Asset Generation
-Generate all image/video/audio assets using `scripts/`. NEVER use placeholder URLs (unsplash, picsum, placeholder.com, via.placeholder, placehold.co, etc.) or external URLs.
+Generate all image/video/audio assets using `scripts/`. Don't use placeholder URLs (unsplash, picsum, placeholder.com, via.placeholder, placehold.co, etc.) or external URLs: they break offline and in production.
 
 1. Parse asset requirements (type, style, spec, usage)
 2. Craft optimized prompts, show to user, confirm before generating
@@ -169,7 +169,7 @@ Generate all image/video/audio assets using `scripts/`. NEVER use placeholder UR
 Follow copywriting frameworks (AIDA, PAS, FAB) to craft all text content. Do NOT use "Lorem ipsum" — write real copy.
 
 ### Phase 5: Build UI
-Scaffold the project and build each section following Design and Motion rules. Integrate generated assets and copy. All `<img>`, `<video>`, `<source>`, and CSS `background-image` MUST reference local assets from Phase 3.
+Scaffold the project and build each section following Design and Motion rules. Integrate generated assets and copy. All `<img>`, `<video>`, `<source>`, and CSS `background-image` reference local assets from Phase 3.
 
 ### Phase 6: Quality Gates
 Run final checklist (see Quality Gates section).
@@ -191,19 +191,19 @@ Adapt dynamically based on user requests.
 ## 1.2 Architecture Conventions
 - **DEPENDENCY VERIFICATION:** Check `package.json` before importing any library. Output install command if missing.
 - **Framework:** React/Next.js. Default to Server Components. Interactive components must be isolated `"use client"` leaf components.
-- **Styling:** Tailwind CSS. Check version in `package.json` — NEVER mix v3/v4 syntax.
-- **ANTI-EMOJI POLICY:** NEVER use emojis anywhere. Use Phosphor or Radix icons only.
+- **Styling:** Tailwind CSS. Check the version in `package.json` and use only that version's syntax; v3 and v4 don't mix.
+- **Icons:** No emojis anywhere. Use Phosphor or Radix icons.
 - **Viewport:** Use `min-h-[100dvh]` not `h-screen`. Use CSS Grid not flex percentage math.
 - **Layout:** `max-w-[1400px] mx-auto` or `max-w-7xl`.
 
 ## 1.3 Design Rules
 | Rule | Directive |
 |------|-----------|
-| Typography | Headlines: `text-4xl md:text-6xl tracking-tighter`. Body: `text-base leading-relaxed max-w-[65ch]`. **NEVER** use Inter — use Geist/Outfit/Satoshi. **NEVER** use Serif on dashboards. |
-| Color | Max 1 accent, saturation < 80%. **NEVER** use AI purple/blue. Stick to one palette. |
-| Layout | **NEVER** use centered heroes when VARIANCE > 4. Force split-screen or asymmetric layouts. |
-| Cards | **NEVER** use generic cards when DENSITY > 7. Use `border-t`, `divide-y`, or spacing. |
-| States | **ALWAYS** implement: Loading (skeleton), Empty, Error, Tactile feedback (`scale-[0.98]`). |
+| Typography | Headlines: `text-4xl md:text-6xl tracking-tighter`. Body: `text-base leading-relaxed max-w-[65ch]`. No Inter: use Geist/Outfit/Satoshi. No serif on dashboards. |
+| Color | Max 1 accent, saturation < 80%. No AI purple/blue. Stick to one palette. |
+| Layout | No centered heroes when VARIANCE > 4: use split-screen or asymmetric layouts. |
+| Cards | No generic cards when DENSITY > 7: use `border-t`, `divide-y`, or spacing. |
+| States | Implement Loading (skeleton), Empty, Error and Tactile feedback (`scale-[0.98]`). |
 | Forms | Label above input. Error below. `gap-2` for input blocks. |
 
 ## 1.4 Anti-Slop Techniques
@@ -271,10 +271,10 @@ When brand styling is active:
 | Hover/focus states | **CSS only** — zero JS cost |
 | Native scroll-driven | **CSS** — `animation-timeline: scroll()` |
 
-**Conflict Rules [MANDATORY]:**
-- NEVER mix GSAP + Framer Motion in same component
-- R3F MUST live in isolated Canvas wrapper
-- ALWAYS lazy-load Lottie, GSAP, Three.js
+**Conflict rules:**
+- Don't mix GSAP and Framer Motion in the same component: both drive the same transforms and fight
+- R3F lives in an isolated Canvas wrapper
+- Lazy-load Lottie, GSAP and Three.js
 
 ## 2.2 Intensity Scale
 
@@ -304,22 +304,22 @@ See `references/motion-recipes.md` for full code. Summary:
 | Layout Morph | Framer | Card-to-modal expansion |
 
 ## 2.4 Performance Rules
-**GPU-only properties (ONLY animate these):** `transform`, `opacity`, `filter`, `clip-path`
+**GPU-only properties (animate only these):** `transform`, `opacity`, `filter`, `clip-path`
 
-**NEVER animate:** `width`, `height`, `top`, `left`, `margin`, `padding`, `font-size` — if you need these effects, use `transform: scale()` or `clip-path` instead.
+**Don't animate** `width`, `height`, `top`, `left`, `margin`, `padding`, `font-size`: they trigger layout on every frame. If you need these effects, use `transform: scale()` or `clip-path` instead.
 
 **Isolation:**
-- Perpetual animations MUST be in `React.memo` leaf components
-- `will-change: transform` ONLY during animation
+- Perpetual animations go in `React.memo` leaf components
+- `will-change: transform` only during animation
 - `contain: layout style paint` on heavy containers
 
 **Mobile:**
-- ALWAYS respect `prefers-reduced-motion`
-- ALWAYS disable parallax/3D on `pointer: coarse`
+- Respect `prefers-reduced-motion`
+- Disable parallax/3D on `pointer: coarse`
 - Cap particles: desktop 800, tablet 300, mobile 100
 - Disable GSAP pin on mobile < 768px
 
-**Cleanup:** Every `useEffect` with GSAP/observers MUST `return () => ctx.revert()`
+**Cleanup:** Every `useEffect` with GSAP/observers returns `() => ctx.revert()`
 
 ## 2.5 Springs & Easings
 
@@ -337,11 +337,10 @@ See `references/motion-recipes.md` for full code. Summary:
 | Elastic | `cubic-bezier(0.34, 1.56, 0.64, 1)` |
 
 ## 2.6 Accessibility
-- ALWAYS wrap motion in `prefers-reduced-motion` check
-- NEVER flash content > 3 times/second (seizure risk)
-- ALWAYS provide visible focus rings (use `outline` not `box-shadow`)
-- ALWAYS add `aria-live="polite"` for dynamically revealed content
-- ALWAYS include pause button for auto-playing animations
+- Never flash content more than 3 times per second (seizure risk)
+- Visible focus rings (`outline`, not `box-shadow`)
+- `aria-live="polite"` on dynamically revealed content
+- A pause button on auto-playing animations
 
 ## 2.7 Dependencies
 
@@ -371,9 +370,9 @@ Env: `MINIMAX_API_KEY` (required).
 
 ## 3.2 Workflow
 1. **Parse:** type, quantity, style, spec, usage
-2. **Craft prompt:** Be specific (composition, lighting, style). **NEVER** include text in image prompts.
-3. **Execute:** Show prompt to user, **MUST confirm before generating**, then run script
-4. **Save:** `<project>/public/assets/{images,videos,audio}/` as `{type}-{descriptor}-{timestamp}.{ext}` — **MUST save locally**
+2. **Craft prompt:** Be specific (composition, lighting, style). Leave text out of image prompts: generators render it garbled.
+3. **Execute:** Show the prompt to the user and wait for confirmation before generating, then run the script
+4. **Save:** `<project>/public/assets/{images,videos,audio}/` as `{type}-{descriptor}-{timestamp}.{ext}` — save locally
 5. **Post-process:** Images → WebP, Videos → ffmpeg compress, Audio → normalize
 6. **Deliver:** File path + code snippet + CSS suggestion
 

package/skills/frontend-dev/references/asset-prompt-guide.md

@@ -7,7 +7,7 @@
 - Include style modifiers: "editorial photography", "3D render", "flat vector illustration"
 - Add technical specs: "4K resolution, sharp focus, shallow depth of field"
 - For web assets: always mention "clean background", "web-optimized", "high contrast for readability"
-- **NEVER** include text in image prompts unless explicitly requested — AI text rendering is unreliable
+- Leave text out of image prompts unless explicitly requested — AI text rendering is unreliable
 
 ## Video Prompts
 

package/skills/frontend-dev/references/motion-recipes.md

@@ -106,7 +106,7 @@ export function PinnedTimeline() {
       });
     }, containerRef);
 
-    return () => ctx.revert(); // CRITICAL: full cleanup
+    return () => ctx.revert(); // full cleanup, or GSAP keeps running after unmount
   }, []);
 
   return (

package/skills/help/SKILL.md

@@ -16,7 +16,7 @@ model: haiku
 - Latest sprint: !`ls -t production/sprints/*.md 2>/dev/null | head -1 || echo 'none'`
 - Session state: !`head -5 production/session-state/active.md 2>/dev/null || echo 'none'`
 
-Read-only skill — reports findings, writes no files. Lightweight orientation; for full gap analysis use `/project-stage-detect`.
+Read-only skill — reports findings, writes no files. Lightweight orientation; for a formal readiness check use `/gate-check`.
 
 ## Step 1: Read Catalog + Find Uncataloged Skills
 
@@ -65,7 +65,6 @@ All required steps complete → "Close to [Current]→[Next] gate. Run `/gate-ch
 
 ## Step 9: Escalation (only if user seems stuck)
 
-- `/project-stage-detect` — full gap analysis
 - `/gate-check` — formal readiness check
 - `/start` — re-orient from scratch
 

package/skills/hotfix/SKILL.md

@@ -50,8 +50,8 @@ Spawn `qa-director` (or `qa-engineer` if that is what is installed) with hotfix
 | QA Verdict | Action |
 |------------|--------|
 | Smoke check sufficient | `/smoke-check` → if PASS, proceed |
-| Targeted QA pass | `/team-qa [affected-system]` → if APPROVED, proceed |
-| Full QA required | `/team-qa sprint` — delays deploy but prevents bad patch |
+| Targeted QA pass | Manual pass on the affected system with the plan from `/qa-plan [affected-system]` → if PASS, proceed |
+| Full QA required | Full manual pass with the plan from `/qa-plan sprint` — delays deploy but prevents bad patch |
 
 Do not skip this gate.
 

package/skills/milestone-review/SKILL.md

@@ -9,7 +9,7 @@ allowed-tools: Read, Glob, Grep, Write, Task, AskUserQuestion
 
 ## Phase 0: Parse Arguments
 
-Extract milestone name (`current` = most recent). Resolve review mode: `--review` arg > `production/review-mode.txt` > default `lean`. See `.claude/docs/director-gates.md`.
+Extract milestone name (`current` = most recent). Resolve review mode: `--review` arg > `production/review-mode.txt` > default `lean`.
 
 ## Phase 1: Load Milestone Data
 

package/skills/patch-notes/SKILL.md

@@ -25,7 +25,7 @@ Read: `production/releases/[version]/changelog.md`, `docs/CHANGELOG.md` entry, `
 
 Check for tone/voice guidance in: `.claude/docs/technical-preferences.md`, `docs/PATCH-NOTES-STYLE.md`, `design/community/tone-guide.md`. Default: player-friendly, non-technical, enthusiastic not hyperbolic, focus on player experience.
 
-Check for template at `docs/patch-notes-template.md` or `.claude/docs/templates/patch-notes-template.md`. If found → use as structure instead of built-in styles.
+Use the project's own template at `docs/patch-notes-template.md` if it exists; otherwise `.claude/docs/templates/release-notes.md` (player-facing: headline, what's new, gameplay changes, fixes). Either one replaces the built-in styles.
 
 ## Phase 3: Categorize and Translate
 

package/skills/playtest-report/SKILL.md

@@ -9,7 +9,7 @@ allowed-tools: Read, Glob, Grep, Write, Task, AskUserQuestion
 
 ## Phase 1: Parse Arguments
 
-Resolve review mode: `--review` arg > `production/review-mode.txt` > default `lean`. See `.claude/docs/director-gates.md`.
+Resolve review mode: `--review` arg > `production/review-mode.txt` > default `lean`.
 
 Modes: `new` → blank template, `analyze [path]` → structured findings from raw notes.
 
@@ -25,7 +25,7 @@ Read raw notes. Cross-reference with design documents. Fill template with struct
 
 | Bucket | Route |
 |--------|-------|
-| Design changes | `/propagate-design-change [path]` for downstream impacts |
+| Design changes | Update the GDD, then `/consistency-check` for downstream impacts |
 | Balance adjustments | `/balance-check [system]` before tuning |
 | Bugs | `/bug-report` to formally track |
 | Polish items | Add to `production/` polish backlog |
@@ -41,5 +41,4 @@ Ask to write to `production/qa/playtests/playtest-[date]-[tester].md`.
 ## Phase 5: Next Steps
 
 - Act on highest-priority finding category first
-- Design changes → `/design-review` on updated GDD
 - Bugs → `/bug-triage` to update priorities

package/skills/playwright-pro/CLAUDE.md

@@ -67,8 +67,8 @@ Leverage Claude Code's built-in capabilities:
 
 - **Large migrations**: Use `/batch` for parallel file-by-file conversion
 - **Post-generation cleanup**: Use `/simplify` after generating a test suite
-- **Debugging sessions**: Use `/debug` alongside `/pw:fix` for trace analysis
-- **Code review**: Use `/review` for general code quality, `/pw:review` for Playwright-specific
+- **Debugging sessions**: Use `/debug` alongside `/playwright-pro fix` for trace analysis
+- **Code review**: Use `/review` for general code quality, `/playwright-pro review` for Playwright-specific
 
 ### Integrations
 

package/skills/playwright-pro/SKILL.md

@@ -1,7 +1,7 @@
 ---
 name: playwright-pro
 description: "Production-grade Playwright testing toolkit. Generate tests, fix flaky failures, migrate from Cypress/Selenium, sync with TestRail, run on BrowserStack. 55 templates, 3 agents, smart reporting."
-argument-hint: ""
+argument-hint: "[browserstack|fix|generate|init|migrate|report|review|testrail] [args]"
 category: "testing"
 user-invocable: true
 allowed-tools: Read, Glob, Grep, Bash, Write, Edit
@@ -13,21 +13,25 @@ Production-grade Playwright testing toolkit for AI coding agents.
 
 > **Note:** This skill may create or modify files in your project. It will ask before writing.
 
+## Sub-commands
+
+`/playwright-pro` with no argument shows this overview. With a sub-command as its first argument, read `skills/<sub-command>/SKILL.md` from this skill's directory and follow it, passing the remaining arguments. Sub-commands: `browserstack`, `fix`, `generate`, `init`, `migrate`, `report`, `review`, `testrail`.
+
 ## Quick Start Workflow
 
 The recommended sequence for most projects:
 
 ```
-1. /pw:init          -> scaffolds config, CI pipeline, and a first smoke test
-2. /pw:generate      -> generates tests from your spec or URL
-3. /pw:review        -> validates quality and flags anti-patterns      <- always run after generate
-4. /pw:fix <test>    -> diagnoses and repairs any failing/flaky tests  <- run when CI turns red
+1. /playwright-pro init          -> scaffolds config, CI pipeline, and a first smoke test
+2. /playwright-pro generate      -> generates tests from your spec or URL
+3. /playwright-pro review        -> validates quality and flags anti-patterns      <- always run after generate
+4. /playwright-pro fix <test>    -> diagnoses and repairs any failing/flaky tests  <- run when CI turns red
 ```
 
 **Validation checkpoints:**
-- After `/pw:generate` -- always run `/pw:review` before committing
-- After `/pw:fix` -- re-run the full suite locally (`npx playwright test`)
-- After `/pw:migrate` -- run `/pw:coverage` to confirm parity before decommissioning old tests
+- After `/playwright-pro generate` -- always run `/playwright-pro review` before committing
+- After `/playwright-pro fix` -- re-run the full suite locally (`npx playwright test`)
+- After `/playwright-pro migrate` -- run both suites and compare test counts and results before decommissioning old tests
 
 ## Golden Rules
 
@@ -55,7 +59,7 @@ The recommended sequence for most projects:
 
 ## What's Included
 
-- **9 skills** with detailed step-by-step instructions
+- **8 skills** with detailed step-by-step instructions
 - **3 specialized agents**: test-architect, test-debugger, migration-planner
 - **55 test templates**: auth, CRUD, checkout, search, forms, dashboard, settings, onboarding, notifications, API, accessibility
 - **2 MCP servers** (TypeScript): TestRail and BrowserStack integrations

package/skills/playwright-pro/agents/migration-planner.md

@@ -2,7 +2,7 @@
 name: migration-planner
 description: >-
   Analyzes Cypress or Selenium test suites and creates a file-by-file
-  migration plan. Invoked by /pw:migrate before conversion starts.
+  migration plan. Invoked by /playwright-pro migrate before conversion starts.
 allowed-tools:
   - Read
   - Grep
@@ -118,4 +118,4 @@ Flag tests that may need manual intervention:
 
 ### Step 7: Return Plan
 
-Return the complete migration plan to `/pw:migrate` for execution.
+Return the complete migration plan to `/playwright-pro migrate` for execution.

package/skills/playwright-pro/agents/test-architect.md

@@ -1,8 +1,7 @@
 ---
 name: test-architect
 description: >-
-  Plans test strategy for complex applications. Invoked by /pw:generate and
-  /pw:coverage when the app has multiple routes, complex state, or requires
+  Plans test strategy for complex applications. Invoked by /playwright-pro generate when the app has multiple routes, complex state, or requires
   a structured test plan before writing tests.
 allowed-tools:
   - Read

package/skills/playwright-pro/agents/test-debugger.md

@@ -2,7 +2,7 @@
 name: test-debugger
 description: >-
   Diagnoses flaky or failing Playwright tests using systematic taxonomy.
-  Invoked by /pw:fix when a test needs deep analysis including running
+  Invoked by /playwright-pro fix when a test needs deep analysis including running
   tests, reading traces, and identifying root causes.
 allowed-tools:
   - Read

package/skills/playwright-pro/hooks/detect-playwright.sh

@@ -20,4 +20,4 @@ fi
 # Count existing test files
 TEST_COUNT=$(find . -name "*.spec.ts" -o -name "*.spec.js" -o -name "*.test.ts" -o -name "*.test.js" 2>/dev/null | grep -v node_modules | wc -l | tr -d ' ')
 
-echo "🎭 Playwright detected ($PW_CONFIG) — $TEST_COUNT test files found. Use /pw: commands for testing workflows."
+echo "🎭 Playwright detected ($PW_CONFIG) — $TEST_COUNT test files found. Use /playwright-pro for testing workflows."

package/skills/playwright-pro/skills/browserstack/SKILL.md

@@ -24,7 +24,7 @@ If not set, inform the user how to get them from [browserstack.com/accounts/sett
 ### 1. Configure for BrowserStack
 
 ```
-/pw:browserstack setup
+/playwright-pro browserstack setup
 ```
 
 Steps:
@@ -96,7 +96,7 @@ export default defineConfig({
 ### 2. Run Tests on BrowserStack
 
 ```
-/pw:browserstack run
+/playwright-pro browserstack run
 ```
 
 Steps:
@@ -113,7 +113,7 @@ Steps:
 ### 3. Get Build Results
 
 ```
-/pw:browserstack results
+/playwright-pro browserstack results
 ```
 
 Steps:
@@ -130,7 +130,7 @@ Steps:
 ### 4. Check Available Browsers
 
 ```
-/pw:browserstack browsers
+/playwright-pro browserstack browsers
 ```
 
 Steps:
@@ -141,7 +141,7 @@ Steps:
 ### 5. Local Testing
 
 ```
-/pw:browserstack local
+/playwright-pro browserstack local
 ```
 
 For testing localhost or staging behind firewall:

package/skills/playwright-pro/skills/migrate/SKILL.md

@@ -48,7 +48,7 @@ Migration Assessment:
 
 ### 3. Set Up Playwright (If Not Present)
 
-Run `/pw:init` first if Playwright isn't configured.
+Run `/playwright-pro init` first if Playwright isn't configured.
 
 ### 4. Convert Files
 

package/skills/playwright-pro/skills/report/SKILL.md

@@ -46,7 +46,7 @@ Check what's configured and route automatically:
 
 | Check | If found | Action |
 |---|---|---|
-| `TESTRAIL_URL` env var | TestRail configured | Push results via `/pw:testrail push` |
+| `TESTRAIL_URL` env var | TestRail configured | Push results via `/playwright-pro testrail push` |
 | `SLACK_WEBHOOK_URL` env var | Slack configured | Post summary to Slack |
 | `.github/workflows/` | GitHub Actions | Results go to PR comment via artifacts |
 | `playwright-report/` | HTML reporter | Open or serve the report |
@@ -98,7 +98,7 @@ curl -X POST "$SLACK_WEBHOOK_URL" \
 
 #### TestRail Push (If Configured)
 
-Invoke `/pw:testrail push` with the JSON results.
+Invoke `/playwright-pro testrail push` with the JSON results.
 
 #### HTML Report