agent-skill-kit 3.9.135

package/.agent/workflows/autopilot.md

@@ -0,0 +1,344 @@
+---
+description: Autonomous multi-agent orchestration — coordinate 3+ specialist agents in parallel with automated verification, conflict resolution, and continuous execution.
+chain: build-web-app
+skills: [lifecycle-orchestrator, execution-reporter, context-engineering, project-planner, studio, problem-checker, idea-storm, design-system, test-architect, security-scanner, knowledge-compiler]
+agents: [orchestrator, assessor, recovery, critic, learner, project-planner, frontend-specialist, backend-specialist, database-architect, security-auditor, test-engineer, devops-engineer, mobile-developer]
+---
+
+# /autopilot - Multi-Agent Command Center
+
+$ARGUMENTS
+
+---
+
+## Purpose
+
+Coordinate 3+ specialist agents for complex multi-domain tasks — planning through parallel execution to verification. **Differs from `/build` (sequential new-app creation) and `/cook` (single-scope targeted tasks) by orchestrating multiple specialists simultaneously.** Uses all 5 meta-agents.
+
+> **MINIMUM 3 SPECIALISTS.** Fewer than 3 ? use direct delegation.
+
+---
+
+## ?? Meta-Agents Integration
+
+| Phase | Agent | Action |
+| ----- | ----- | ------ |
+| **Pre-Flight** | `assessor` | Evaluate plan risk level and knowledge-compiler context |
+| **Execution** | `orchestrator` / `critic` | Coordinate parallel execution and resolve agent conflicts |
+| **Safety** | `recovery` | Save state and recover from execution failures |
+| **Post-Autopilot** | `learner` | Log execution metrics and success/failure patterns |
+
+```
+Flow:
+orchestrator.init() ? assessor.evaluate(plan)
+       ?
+recovery.save(state) ? execute phases in parallel
+       ?
+conflict? ? critic.arbitrate(safety > correctness > convenience)
+       ?
+failure? ? recovery.restore() ? learner.log(failure)
+       ?
+success ? learner.log(patterns)
+```
+
+---
+
+## ?? MANDATORY: Multi-Agent Orchestration Protocol
+
+### Phase 1: Pre-flight & knowledge-compiler Context
+
+> **Rule 0.5-K:** knowledge-compiler pattern check.
+
+1. Read `.agent/skills/knowledge-compiler/patterns/` for past failures before proceeding.
+2. Trigger `recovery` agent to run Checkpoint (`git commit -m "chore(checkpoint): pre-autopilot"`).
+
+### Phase 2: Planning & Agent Selection
+
+| Field | Value |
+|-------|-------|
+| **INPUT** | $ARGUMENTS (user request — feature/app description) |
+| **OUTPUT** | PLAN.md with task breakdown, agent assignments, dependency graph |
+| **AGENTS** | `project-planner`, `assessor` |
+| **SKILLS** | `project-planner`, `idea-storm`, `context-engineering` |
+
+// turbo — telemetry: phase-2-planning
+
+1. Identify required domains ? select agents:
+
+| Domain | Agent |
+|--------|-------|
+| Backend/API | `nodejs-pro` |
+| Frontend/UI | `react-pro` |
+| Database | `data-modeler` |
+| Security | `security-scanner` |
+| Testing | `test-architect` |
+| DevOps | `cicd-pipeline` |
+| Mobile | `mobile-developer` |
+
+2. Select minimum 3 agents
+3. Create PLAN.md with assignments and parallel groups
+
+> ?? Only `project-planner` during planning.
+
+**? CHECKPOINT: User approval required before Phase 3**
+
+### Phase 3: Pre-Flight Safety
+
+| Field | Value |
+|-------|-------|
+| **INPUT** | Approved PLAN.md |
+| **OUTPUT** | Risk assessment + state checkpoint |
+| **AGENTS** | `assessor`, `recovery` |
+| **SKILLS** | `lifecycle-orchestrator` |
+
+// turbo — telemetry: phase-3-safety
+
+1. `assessor` evaluates plan risk:
+
+| Risk Level | Criteria | Action |
+|------------|----------|--------|
+| **Low** | Read-only, npm scripts, lint | Auto-proceed |
+| **Medium** | File creation, config changes | Proceed with checkpoint |
+| **High** | DB migrations, multi-file refactor | Extra checkpoints |
+| **Critical** | Auth changes, production deploy | Require explicit user confirmation |
+
+2. `recovery` creates state checkpoint for all affected files
+
+### Phase 4: Design System (UI Apps Only)
+
+| Field | Value |
+|-------|-------|
+| **INPUT** | Approved PLAN.md (if app has UI) |
+| **OUTPUT** | Design tokens (colors, typography, effects) |
+| **AGENTS** | `react-pro`, `orchestrator` |
+| **SKILLS** | `studio`, `design-system` |
+
+> **Skip this phase** if the task has no UI component.
+
+// turbo — telemetry: phase-4-studio-search
+```bash
+npx cross-env OTEL_SERVICE_NAME="workflow:autopilot" TRACE_ID="$TRACE_ID" node .agent/skills/studio/scripts/search.ts "<app_type> <style> <keywords>" --design-system -p "<Project Name>"
+```
+
+Apply design tokens (colors, typography, effects) before building components.
+
+### Phase 5: Parallel Execution
+
+| Field | Value |
+|-------|-------|
+| **INPUT** | Approved plan + design system tokens (if applicable) |
+| **OUTPUT** | All project artifacts: routes, components, schemas, configs |
+| **AGENTS** | `orchestrator`, `critic`, 3+ domain specialists per plan |
+| **SKILLS** | Per agent specialization |
+
+// turbo — telemetry: phase-5-execution
+
+```mermaid
+graph TD
+    A[Approved Plan] --> B[Foundation]
+    B --> B1[database-architect]
+    B --> B2[security-auditor]
+    B1 --> C[Core]
+    B2 --> C
+    C --> C1[backend-specialist]
+    C --> C2[frontend-specialist]
+    C1 --> D[Polish]
+    C2 --> D
+    D --> D1[test-engineer]
+    D --> D2[devops-engineer]
+    D1 --> E[Verification]
+    D2 --> E
+```
+
+| Parallel Group | Agents | Runs After |
+|----------------|--------|------------|
+| **Foundation** | `data-modeler`, `security-scanner` | Plan approved |
+| **Core** | `nodejs-pro`, `react-pro` | Foundation complete |
+| **Polish** | `test-architect`, `cicd-pipeline` | Core complete |
+
+**Agent Selection:** Match task type to specialists. Web ? frontend + backend + test. API ? backend + security + test. Full Stack ? all core agents.
+
+**Context Passing (MANDATORY):** Every sub-agent must receive: original request, decisions made, previous work summary, current plan link, and specific task.
+
+> ?? Invoking agent without context = wrong assumptions!
+
+### Phase 6: Verification & Exit Gate
+
+| Field | Value |
+|-------|-------|
+| **INPUT** | All artifacts from Phase 5 |
+| **OUTPUT** | Verification report: tests, lint, types, security scan |
+| **AGENTS** | `test-architect`, `learner` |
+| **SKILLS** | `test-architect`, `problem-checker`, `security-scanner`, `knowledge-compiler` |
+
+// turbo — telemetry: phase-6-test
+```bash
+npx cross-env OTEL_SERVICE_NAME="workflow:autopilot" TRACE_ID="$TRACE_ID" npm test
+```
+
+// turbo — telemetry: phase-6-lint-typecheck
+```bash
+npx cross-env OTEL_SERVICE_NAME="workflow:autopilot" TRACE_ID="$TRACE_ID" npm run lint; npx cross-env OTEL_SERVICE_NAME="workflow:autopilot" TRACE_ID="$TRACE_ID" npx tsc --noEmit
+```
+
+**Exit Gate — ALL must pass:**
+
+| Check | Target | How to Verify |
+|-------|--------|---------------|
+| Agent Count | = 3 | Count unique agents invoked |
+| IDE Problems | 0 | `@[current_problems]` check |
+| Security Scan | Pass | `security-scanner` output |
+| Lint/Types | 0 errors | ESLint + TypeScript |
+| Tests | All passing | `npm test` output |
+| Preview Running | Yes | `npm run dev` active |
+
+**Exit Decision:** IDE problems > 0 ? auto-fix or notify. Security failed ? block. All deliverables complete ? generate report.
+
+---
+
+## ?? Auto-Execution Policy
+
+**CONTINUOUS EXECUTION** after plan approval. All commands auto-run. Only stop for: blocking errors, decision forks, plan completion.
+
+---
+
+## ? MANDATORY: Problem Verification Before Completion
+
+> **CRITICAL:** This check MUST be performed before any `notify_user` or task completion.
+
+### Check @[current_problems]
+
+```
+1. Read @[current_problems] from IDE
+2. If errors/warnings > 0:
+   a. Auto-fix: imports, types, lint errors
+   b. Re-check @[current_problems]
+   c. If still > 0 ? STOP ? Notify user
+3. If count = 0 ? Proceed to completion
+```
+
+### Auto-Fixable
+
+| Type | Fix |
+|------|-----|
+| Missing import | Add import statement |
+| JSX namespace | Import from 'react' |
+| Unused variable | Remove or prefix `_` |
+| Lint errors | Run eslint --fix |
+| Type mismatch | Fix type annotation |
+
+> **Rule:** Never mark complete with errors in `@[current_problems]`.
+
+---
+
+## ?? Rollback & Recovery
+
+If the Exit Gates fail and cannot be resolved automatically:
+1. Restore to pre-autopilot checkpoint (`git checkout -- .` or `git stash pop`).
+2. Log failure via `learner` meta-agent.
+3. Notify user with failure context and recovery options.
+
+---
+
+## Output Format
+
+```markdown
+## ?? Autopilot Report
+
+### Mission
+[Original task summary]
+
+### Agent Coordination
+
+| Agent | Task | Duration | Status |
+|-------|------|----------|--------|
+| `project-planner` | Task breakdown | 2m | ? Complete |
+| `data-modeler` | Schema design | 3m | ? Complete |
+| `nodejs-pro` | API routes | 5m | ? Complete |
+| `react-pro` | UI components | 7m | ? Complete |
+| `test-architect` | E2E tests | 4m | ? Complete |
+
+### Execution Metrics
+
+| Metric | Target | Actual | Status |
+|--------|--------|--------|--------|
+| Agents invoked | = 3 | [X] | ?/? |
+| IDE problems | 0 | [X] | ?/? |
+| Total execution | < 5min | [X]m | ?/? |
+| Auto-fix rate | > 85% | [X]% | ?/? |
+
+### Deliverables
+- [x] PLAN.md created
+- [x] Database schema
+- [x] API endpoints ([X] routes)
+- [x] UI components ([X] pages)
+- [x] Tests ([X] cases passing)
+- [x] Preview running at localhost:3000
+
+### Next Steps
+- [ ] Review generated code
+- [ ] Test user flows
+- [ ] `/launch` when ready to deploy
+```
+
+---
+
+## Examples
+
+```
+/autopilot build a SaaS dashboard with analytics and auth
+/autopilot create REST API with rate limiting and comprehensive tests
+/autopilot refactor monolith to microservices architecture
+/autopilot add real-time notifications to existing app
+/autopilot security audit + fix vulnerabilities + add test coverage
+```
+
+---
+
+## Key Principles
+
+- **Minimum 3 agents** — autopilot means multi-specialist coordination, not single-agent delegation
+- **Plan first, execute after** — no execution without approved PLAN.md
+- **Parallel by default** — independent agents run simultaneously to reduce total time
+- **Context passing mandatory** — every sub-agent receives full context (request, decisions, prior work)
+- **Exit gate enforced** — IDE problems = 0, security scan passed, all tests green before completion
+
+---
+
+## ?? Workflow Chain
+
+**Skills Loaded (11):**
+
+- `lifecycle-orchestrator` - Multi-phase execution with checkpoint/restore
+- `execution-reporter` - Agent routing and skill loading transparency
+- `context-engineering` - Token optimization and agent architecture
+- `project-planner` - Task breakdown and plan generation
+- `studio` - Design system generation for UI apps
+- `problem-checker` - IDE error detection and auto-fix
+- `idea-storm` - Brainstorming and alternative analysis
+- `design-system` - UI/UX standards and theme tokens
+- `test-architect` - Testing strategy and suite generation
+- `security-scanner` - Automated vulnerability detection
+- `knowledge-compiler` - Learning and logging execution patterns
+
+```mermaid
+graph LR
+    A["/think"] --> B["/plan"]
+    B --> C["/autopilot"]
+    C --> D["/launch"]
+    style C fill:#10b981
+```
+
+| After /autopilot | Run | Purpose |
+|-----------------|-----|---------|
+| All phases complete | `/launch` | Deploy to production |
+| Issues found | `/diagnose` | Root cause investigation |
+| Need more features | `/build` | Enhance existing app |
+| Need test coverage | `/validate` | Comprehensive test suite |
+
+**Handoff to /launch:**
+
+```markdown
+? Autopilot complete! [X] agents finished, [Y] tests passing. Preview running at localhost:3000.
+Run `/launch` when ready to deploy to production.
+```

package/.agent/workflows/build.md

@@ -0,0 +1,338 @@
+---
+description: Full-stack application factory — transform natural language descriptions into production-ready apps with multi-agent coordination, smart stack defaults, and automated verification.
+chain: build-web-app
+skills: [app-scaffold, react-pro, nextjs-pro, project-planner, studio, design-system, code-craft, data-modeler, api-architect, test-architect, idea-storm, nodejs-pro, problem-checker, context-engineering, knowledge-compiler]
+agents: [orchestrator, assessor, recovery, critic, learner, project-planner, database-architect, backend-specialist, frontend-specialist, test-engineer, mobile-developer, devops-engineer, security-auditor]
+---
+
+# /build - Application Factory
+
+$ARGUMENTS
+
+---
+
+## Purpose
+
+Ship production-ready applications from natural language descriptions by coordinating 4+ specialist agents with automated verification. **Differs from `/autopilot` (fully autonomous multi-domain orchestration) and `/cook` (targeted single-scope tasks) by focusing on new application creation from scratch with guided requirements discovery.** Uses `project-planner` for architecture, `data-modeler` + `nodejs-pro` + `react-pro` for parallel implementation, and `test-architect` for verification.
+
+---
+
+## ?? Meta-Agents Integration
+
+| Phase | Agent | Action |
+| ----- | ----- | ------ |
+| **Pre-Flight** | `assessor` | Evaluate build risk level and knowledge-compiler context |
+| **Execution** | `orchestrator` / `critic` | Coordinate parallel execution and resolve schema/API conflicts |
+| **Safety** | `recovery` | Save state and recover from build failures |
+| **Post-Build** | `learner` | Log build architecture and implementation patterns |
+
+```
+Flow:
+assessor.evaluate(risk) ? recovery.save(state)
+       ?
+orchestrator.init() ? assign agents to parallel groups
+       ?
+conflict? ? critic.arbitrate(safety > correctness)
+       ?
+failure? ? recovery.restore() ? learner.log(failure)
+       ?
+success ? learner.log(patterns)
+```
+
+---
+
+## ?? MANDATORY: Build Pipeline
+
+### Phase 1: Pre-flight & knowledge-compiler Context
+
+> **Rule 0.5-K:** knowledge-compiler pattern check.
+
+1. Read `.agent/skills/knowledge-compiler/patterns/` for past failures before proceeding.
+2. Trigger `recovery` agent to run Checkpoint (`git commit -m "chore(checkpoint): pre-build"`).
+
+### Phase 2: Requirements Discovery
+
+| Field | Value |
+|-------|-------|
+| **INPUT** | $ARGUMENTS (app description — what to build) |
+| **OUTPUT** | Requirements document: app type, users, core features, stack |
+| **AGENTS** | `project-planner`, `assessor` |
+| **SKILLS** | `idea-storm`, `project-planner`, `context-engineering` |
+
+// turbo — telemetry: phase-2-requirements
+
+1. If requirements are incomplete, ask clarifying questions:
+
+```
+1. What TYPE of app? (web/mobile/api/cli)
+2. Who are the USERS? (consumers/business/developers)
+3. What are the CORE FEATURES? (list 3-5 must-haves)
+4. What STACK preferences? (or let me choose)
+5. What is the TIMELINE? (MVP/full product)
+```
+
+2. Select tech stack based on app type:
+
+| App Type | Recommended Stack | Agents Invoked |
+|----------|------------------|----------------|
+| Web SaaS | Next.js + Prisma + PostgreSQL | frontend, backend, database |
+| E-commerce | Next.js + Stripe + Supabase | frontend, backend, security |
+| Mobile | React Native + Expo + Firebase | mobile, backend, test |
+| API | Hono + Prisma + Railway | backend, database, devops |
+| Dashboard | Next.js + Chart.js | frontend, backend |
+
+3. Smart defaults (when user doesn't specify):
+
+| Aspect | Default |
+|--------|---------|
+| Framework | Next.js 15 |
+| Styling | Tailwind + shadcn/ui |
+| Database | PostgreSQL (Supabase) |
+| Auth | Clerk |
+| Deployment | Vercel |
+| Testing | Vitest + Playwright |
+
+### Phase 3: Planning & Design
+
+| Field | Value |
+|-------|-------|
+| **INPUT** | Requirements from Phase 2 |
+| **OUTPUT** | PLAN.md with task breakdown, agent assignments, file structure |
+| **AGENTS** | `project-planner` |
+| **SKILLS** | `project-planner`, `code-craft` |
+
+// turbo — telemetry: phase-3-planning
+
+1. Create PLAN.md with:
+   - Task breakdown and agent assignments
+   - File structure blueprint
+   - Dependency graph between agents
+   - Parallel execution groups
+2. `assessor` evaluates risk level of the build plan
+3. `recovery` saves current project state (if existing directory)
+
+**? CHECKPOINT: User approval required before Phase 4**
+
+### Phase 4: Design System (UI Apps Only)
+
+| Field | Value |
+|-------|-------|
+| **INPUT** | Approved PLAN.md (for apps with UI) |
+| **OUTPUT** | Design tokens: colors, typography, effects |
+| **AGENTS** | `react-pro`, `orchestrator` |
+| **SKILLS** | `studio`, `design-system` |
+
+> **Skip this phase** if the app has no UI (API-only, CLI).
+
+// turbo — telemetry: phase-4-studio-search
+```bash
+npx cross-env OTEL_SERVICE_NAME="workflow:build" TRACE_ID="$TRACE_ID" node .agent/skills/studio/scripts/search.ts "<app_type> <style> <keywords>" --design-system -p "<Project Name>"
+```
+
+### Phase 5: Parallel Build
+
+| Field | Value |
+|-------|-------|
+| **INPUT** | Approved plan + design system tokens |
+| **OUTPUT** | Complete application: schema, API routes, UI components, tests |
+| **AGENTS** | `orchestrator`, `critic`, `data-modeler`, `nodejs-pro`, `react-pro`, `test-architect` |
+| **SKILLS** | `data-modeler`, `nodejs-pro`, `api-architect`, `code-craft`, `test-architect` |
+
+// turbo — telemetry: phase-5-build
+
+```mermaid
+graph TD
+    A[Approved Plan] --> B[Foundation]
+    B --> B1[database-architect]
+    B --> B2[security-auditor]
+    B1 --> C[Core]
+    B2 --> C
+    C --> C1[backend-specialist]
+    C --> C2[frontend-specialist]
+    C1 --> D[Polish]
+    C2 --> D
+    D --> D1[test-engineer]
+    D1 --> E[Preview]
+```
+
+| Parallel Group | Agents | Task |
+|----------------|--------|------|
+| **Foundation** | `data-modeler`, `security-scanner` | Schema + auth setup |
+| **Core** | `nodejs-pro`, `react-pro` | API routes + UI components |
+| **Polish** | `test-architect` | Tests + integration |
+
+### Phase 6: Verification & Preview
+
+| Field | Value |
+|-------|-------|
+| **INPUT** | All artifacts from Phase 5 |
+| **OUTPUT** | Running preview + verification report |
+| **AGENTS** | `test-architect`, `learner` |
+| **SKILLS** | `test-architect`, `problem-checker`, `knowledge-compiler` |
+
+// turbo — telemetry: phase-6-test
+```bash
+npx cross-env OTEL_SERVICE_NAME="workflow:build" TRACE_ID="$TRACE_ID" npm test
+```
+
+// turbo — telemetry: phase-6-lint-typecheck
+```bash
+npx cross-env OTEL_SERVICE_NAME="workflow:build" TRACE_ID="$TRACE_ID" npm run lint; npx cross-env OTEL_SERVICE_NAME="workflow:build" TRACE_ID="$TRACE_ID" npx tsc --noEmit
+```
+
+// turbo — telemetry: phase-6-preview
+```bash
+npx cross-env OTEL_SERVICE_NAME="workflow:build" TRACE_ID="$TRACE_ID" npm run dev
+```
+
+**Exit Gate — ALL must pass:**
+
+| Check | Target | How to Verify |
+|-------|--------|---------------|
+| IDE Problems | 0 | `@[current_problems]` |
+| Compilation | Pass | TypeScript check |
+| Tests | All passing | npm test |
+| Preview | Running | localhost:3000 responding |
+| Schema | Valid | Prisma validate (if applicable) |
+
+---
+
+## ? MANDATORY: Problem Verification Before Completion
+
+> **CRITICAL:** This check MUST be performed before any `notify_user` or task completion.
+
+### Check @[current_problems]
+
+```
+1. Read @[current_problems] from IDE
+2. If errors/warnings > 0:
+   a. Auto-fix: imports, types, lint errors
+   b. Re-check @[current_problems]
+   c. If still > 0 ? STOP ? Notify user
+3. If count = 0 ? Proceed to completion
+```
+
+### Auto-Fixable
+
+| Type | Fix |
+|------|-----|
+| Missing import | Add import statement |
+| JSX namespace | Import from 'react' |
+| Unused variable | Remove or prefix `_` |
+| Lint errors | Run eslint --fix |
+
+> **Rule:** Never mark complete with errors in `@[current_problems]`. **NEVER mark complete without a working preview.**
+
+---
+
+## ?? Rollback & Recovery
+
+If the Exit Gates fail and cannot be resolved automatically:
+1. Restore to pre-build checkpoint (`git checkout -- .` or `git stash pop`).
+2. Log failure via `learner` meta-agent.
+3. Notify user with failure context and recovery options.
+
+---
+
+## Output Format
+
+```markdown
+## ??? Building: [App Name]
+
+### Stack Decision
+
+| Layer | Choice | Reason |
+|-------|--------|--------|
+| Frontend | Next.js 15 | SSR, App Router, Vercel deploy |
+| Backend | Hono + Prisma | Type-safe, edge-ready |
+| Database | PostgreSQL | Relational, Supabase hosting |
+| Auth | Clerk | Fast integration, social login |
+| Styling | Tailwind + shadcn/ui | Rapid UI development |
+
+### Agent Coordination
+
+| Agent | Task | Status |
+|-------|------|--------|
+| `project-planner` | Architecture plan | ? Complete |
+| `data-modeler` | Schema design | ? Complete |
+| `nodejs-pro` | API routes | ? Complete |
+| `react-pro` | UI components | ? Complete |
+| `test-architect` | E2E tests | ? Complete |
+
+### Preview
+
+?? http://localhost:3000
+
+### Next Steps
+
+- [ ] Review the generated code
+- [ ] Test core user flows
+- [ ] Run `/validate` for comprehensive tests
+- [ ] Deploy when ready: `/launch`
+```
+
+---
+
+## Examples
+
+```
+/build SaaS dashboard with user analytics and auth
+/build e-commerce store with Stripe payments and product catalog
+/build blog engine with MDX content and CMS
+/build real-time chat app with WebSocket and message history
+/build REST API for mobile app with auth and rate limiting
+```
+
+---
+
+## Key Principles
+
+- **Requirements first** — discover what to build before building, ask clarifying questions
+- **Smart stack defaults** — sensible tech choices unless user specifies otherwise
+- **Parallel execution** — independent agents run simultaneously for faster delivery
+- **Working preview required** — never mark complete without a running dev server
+- **Exit gate enforced** — IDE problems = 0, tests passing, types valid before completion
+
+---
+
+## ?? Workflow Chain
+
+**Skills Loaded (15):**
+
+- `code-craft` - Full-stack scaffolding from natural language
+- `react-pro` - React application development
+- `nextjs-pro` - Next.js framework expertise
+- `project-planner` - Task breakdown and architecture planning
+- `studio` - Design system generation for UI apps
+- `design-system` - UI/UX design tokens and components
+- `code-craft` - Coding standards and best practices
+- `data-modeler` - Database schema design
+- `api-architect` - API design and endpoint patterns
+- `test-architect` - Testing strategies and coverage
+- `idea-storm` - Brainstorming and alternative analysis
+- `nodejs-pro` - Node.js backend development
+- `context-engineering` - Codebase parsing and framework detection
+- `problem-checker` - IDE error detection and auto-fix
+- `knowledge-compiler` - Learning and logging execution patterns
+
+```mermaid
+graph LR
+    A["/plan"] --> B["/build"]
+    B --> C["/validate"]
+    C --> D["/launch"]
+    style B fill:#10b981
+```
+
+| After /build | Run | Purpose |
+|-------------|-----|---------|
+| Build complete | `/validate` | Run comprehensive test suite |
+| Needs preview | `/stage` | Start development environment |
+| Ready to deploy | `/launch` | Deploy to production |
+| Found bugs | `/diagnose` | Root cause investigation |
+
+**Handoff to /validate:**
+
+```markdown
+? Build complete! Preview running at localhost:3000.
+[X] agents coordinated, [Y] files created. Run `/validate` for full test suite.
+```