@hustle-together/api-dev-tools 3.11.1 → 3.12.1

package/.skills/README.md

@@ -1,6 +1,6 @@
 # API Development Tools - Agent Skills
 
-**Version:** 3.11.0
+**Version:** 1.0.0
 **Standard:** [Agent Skills Open Format](https://agentskills.io)
 **Platform:** Cross-platform (Claude Code, VS Code, Cursor, ChatGPT, GitHub Copilot)
 
@@ -32,55 +32,55 @@ npx @hustle-together/api-dev-tools --scope=project
 
 ### API Development (6 skills)
 
-| Skill | Usage | Description |
-|-------|-------|-------------|
-| **api-create** | `/api-create [endpoint]` | Complete 13-phase API workflow with interview, research, TDD, verification |
-| **api-interview** | `/api-interview [endpoint]` | Structured requirements gathering with questions FROM research findings |
-| **api-research** | `/api-research [library]` | Adaptive documentation research with 7-day caching |
-| **api-verify** | `/api-verify [endpoint]` | Re-research after tests pass to catch implementation gaps |
-| **api-env** | `/api-env [endpoint]` | Check API keys and environment variables |
-| **api-status** | `/api-status [endpoint]` | Track progress through 13 phases |
+| Skill             | Usage                       | Description                                                                |
+| ----------------- | --------------------------- | -------------------------------------------------------------------------- |
+| **api-create**    | `/api-create [endpoint]`    | Complete 14-phase API workflow with interview, research, TDD, verification |
+| **api-interview** | `/api-interview [endpoint]` | Structured requirements gathering with questions FROM research findings    |
+| **api-research**  | `/api-research [library]`   | Adaptive documentation research with 7-day caching                         |
+| **api-verify**    | `/api-verify [endpoint]`    | Re-research after tests pass to catch implementation gaps                  |
+| **api-env**       | `/api-env [endpoint]`       | Check API keys and environment variables                                   |
+| **api-status**    | `/api-status [endpoint]`    | Track progress through 14 phases                                           |
 
 ### TDD Workflow (4 skills)
 
-| Skill | Usage | Description |
-|-------|-------|-------------|
-| **red** | `/red` | Write ONE failing test (defines success before implementation) |
-| **green** | `/green` | Minimal implementation to pass tests (no over-engineering) |
-| **refactor** | `/refactor` | Clean up code while keeping tests green |
-| **cycle** | `/cycle [description]` | Complete Red → Green → Refactor loop |
+| Skill        | Usage                  | Description                                                    |
+| ------------ | ---------------------- | -------------------------------------------------------------- |
+| **red**      | `/red`                 | Write ONE failing test (defines success before implementation) |
+| **green**    | `/green`               | Minimal implementation to pass tests (no over-engineering)     |
+| **refactor** | `/refactor`            | Clean up code while keeping tests green                        |
+| **cycle**    | `/cycle [description]` | Complete Red → Green → Refactor loop                           |
 
 ### Planning & Analysis (3 skills)
 
-| Skill | Usage | Description |
-|-------|-------|-------------|
-| **plan** | `/plan [feature]` | Create implementation plan with PRD-style discovery |
-| **gap** | `/gap` | Analyze code vs requirements for missing pieces |
-| **issue** | `/issue [url]` | Create TDD plan from GitHub issue |
+| Skill     | Usage             | Description                                         |
+| --------- | ----------------- | --------------------------------------------------- |
+| **plan**  | `/plan [feature]` | Create implementation plan with PRD-style discovery |
+| **gap**   | `/gap`            | Analyze code vs requirements for missing pieces     |
+| **issue** | `/issue [url]`    | Create TDD plan from GitHub issue                   |
 
 ### Git Operations (5 skills)
 
-| Skill | Usage | Description |
-|-------|-------|-------------|
-| **commit** | `/commit` | Semantic commit with co-author attribution |
-| **pr** | `/pr` | Create pull request with summary and test plan |
-| **busycommit** | `/busycommit` | Multiple atomic commits for complex changesets |
-| **worktree-add** | `/worktree-add [branch/issue]` | Add git worktree from branch or issue |
-| **worktree-cleanup** | `/worktree-cleanup` | Clean up merged worktrees |
+| Skill                | Usage                          | Description                                    |
+| -------------------- | ------------------------------ | ---------------------------------------------- |
+| **commit**           | `/commit`                      | Semantic commit with co-author attribution     |
+| **pr**               | `/pr`                          | Create pull request with summary and test plan |
+| **busycommit**       | `/busycommit`                  | Multiple atomic commits for complex changesets |
+| **worktree-add**     | `/worktree-add [branch/issue]` | Add git worktree from branch or issue          |
+| **worktree-cleanup** | `/worktree-cleanup`            | Clean up merged worktrees                      |
 
 ### Workflow Utilities (5 skills)
 
-| Skill | Usage | Description |
-|-------|-------|-------------|
-| **spike** | `/spike` | Exploratory coding before formal TDD |
-| **tdd** | `/tdd` | Remind agent about TDD practices |
-| **beepboop** | `/beepboop` | Transparent AI attribution markers |
-| **summarize** | `/summarize` | Conversation progress summary |
-| **add-command** | `/add-command` | Guide for creating new skills |
+| Skill           | Usage          | Description                          |
+| --------------- | -------------- | ------------------------------------ |
+| **spike**       | `/spike`       | Exploratory coding before formal TDD |
+| **tdd**         | `/tdd`         | Remind agent about TDD practices     |
+| **beepboop**    | `/beepboop`    | Transparent AI attribution markers   |
+| **summarize**   | `/summarize`   | Conversation progress summary        |
+| **add-command** | `/add-command` | Guide for creating new skills        |
 
 ## 🏗️ Architecture
 
-### 13-Phase Workflow
+### 14-Phase Workflow
 
 ```
 Phase 1:  DISAMBIGUATION     - Clarify ambiguous terms before research
@@ -93,21 +93,24 @@ Phase 7:  ENVIRONMENT        - Verify API keys exist
 Phase 8:  TDD RED            - Write failing tests from schema
 Phase 9:  TDD GREEN          - Minimal implementation to pass
 Phase 10: VERIFY             - Re-research docs, compare to implementation
-Phase 11: TDD REFACTOR       - Clean up code while tests pass
-Phase 12: DOCUMENTATION      - Update manifests, cache research
-Phase 13: COMPLETION         - Final verification, commit
+Phase 11: CODE REVIEW        - Greptile AI review (bugs, security, performance)
+Phase 12: TDD REFACTOR       - Fix review issues, clean up code
+Phase 13: DOCUMENTATION      - Update manifests, cache research
+Phase 14: COMPLETION         - Final verification, commit
 ```
 
 ### Loop-Back Architecture
 
 Every verification phase can loop back if not successful:
+
 - Phase 1: Loop if disambiguation unclear
 - Phase 3: Loop if more research needed
 - Phase 4: Loop if interview incomplete
 - Phase 6: Loop if schema incorrect
 - Phase 8: Loop if test plan needs changes
 - Phase 10: Loop if gaps found in implementation
-- Phase 12: Loop if documentation incomplete
+- Phase 11: Loop if code review finds issues to fix
+- Phase 13: Loop if documentation incomplete
 
 ### State Tracking
 
@@ -152,22 +155,22 @@ Documentation cached in `.claude/research/` with 7-day freshness:
 
 ### Optional but Recommended
 
-- **Enforcement Hooks**: 18 Python scripts that enforce workflow rules
+- **Enforcement Hooks**: 22 Python scripts that enforce workflow rules
 - **Settings.json**: Hook registration for lifecycle events
 
-## 🔒 Enforcement Hooks (Optional)
+## 🔒 Enforcement Hooks (22 Total)
 
-For full workflow enforcement, install hooks to `.claude/hooks/`:
+For full workflow enforcement, install hooks to `hooks/`:
 
 ### Hook Types
 
-| Event | Hooks | Purpose |
-|-------|-------|---------|
-| **SessionStart** | session-startup.py | Inject state context at session start |
-| **UserPromptSubmit** | enforce-external-research.py | Detect API terms, require research |
-| **PreToolUse** | enforce-research.py<br>enforce-interview.py<br>verify-implementation.py | Block writes until research/interview done<br>Inject interview decisions<br>Require test file before route |
-| **PostToolUse** | track-tool-use.py<br>periodic-reground.py<br>verify-after-green.py | Log research, count turns<br>Re-ground every 7 turns<br>Trigger Phase 10 after test pass |
-| **Stop** | api-workflow-check.py | Block if phases incomplete |
+| Event                | Hooks                                                                                                                          | Purpose                                                                           |
+| -------------------- | ------------------------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------- |
+| **SessionStart**     | session-startup.py                                                                                                             | Inject state context at session start                                             |
+| **UserPromptSubmit** | enforce-external-research.py                                                                                                   | Detect API terms, require research                                                |
+| **PreToolUse**       | enforce-research.py, enforce-interview.py, verify-implementation.py, enforce-component-type-confirm.py, notify-input-needed.py | Block writes until phases complete, inject decisions, confirm component type      |
+| **PostToolUse**      | track-tool-use.py, periodic-reground.py, verify-after-green.py, notify-phase-complete.py, track-token-usage.py                 | Log research, re-ground every 7 turns, trigger verification, notify, track tokens |
+| **Stop**             | api-workflow-check.py                                                                                                          | Block if phases incomplete                                                        |
 
 ### What Hooks Enforce
 
@@ -194,7 +197,7 @@ For full workflow enforcement, install hooks to `.claude/hooks/`:
 # Fully automated workflow
 /api-create my-endpoint
 
-# Follows all 13 phases automatically
+# Follows all 14 phases automatically
 # Asks for user input at each checkpoint
 # Loops back if verification fails
 # Creates tests, implementation, docs
@@ -282,10 +285,9 @@ See [CONTRIBUTING.md](../../CONTRIBUTING.md) for guidelines
 
 - [Agent Skills Specification](https://agentskills.io/specification)
 - [Claude Code Documentation](https://code.claude.com/docs)
-- [Enhancement Roadmap](../../ENHANCEMENT_ROADMAP_v3.11.0.md)
-- [Skills Migration Checklist](../../SKILLS_MIGRATION_CHECKLIST.md)
+- [Future Roadmap](../../ROADMAP.md)
 
 ---
 
 **Built with ❤️ by Hustle Together**
-*Empowering developers with interview-driven, research-first API development*
+_Empowering developers with interview-driven, research-first API development_

package/.skills/_shared/settings.json

@@ -90,4 +90,4 @@
       }
     ]
   }
-}
+}

package/.skills/add-command/SKILL.md

@@ -12,8 +12,10 @@ allowed-tools: WebSearch WebFetch mcp__context7 mcp__github AskUserQuestion Read
 ---
 
 ---
+
 description: Guide for creating new slash commands
 argument-hint: <command-name> <description>
+
 ---
 
 ## General Guidelines
@@ -67,7 +69,7 @@ Bash command output: (exclamation)git status(backticks)
 **Bash Commands (exclamation prefix)**: Limited to current working directory only.
 
 - ✅ Works: `! + backtick + git status + backtick` (in project dir)
-- ❌ Blocked: `! + backtick + ls /outside/project + backtick` (outside project)  
+- ❌ Blocked: `! + backtick + ls /outside/project + backtick` (outside project)
 - ❌ Blocked: `! + backtick + pwd + backtick` (if referencing dirs outside project)
 
 **File References (`@` prefix)**: No directory restrictions.
@@ -103,6 +105,7 @@ Compare @src/old.js with @src/new.js and explain differences
 ---
 allowed-tools: Bash(git status:*), Bash(git branch:*), Bash(git log:*)
 ---
+
 Current status: (!)git status(`)
 Current branch: (!)git branch --show-current(`)
 Recent commits: (!)git log --oneline -5(`)
@@ -136,7 +139,8 @@ echo "Ask GPT-5 about: ＄ARGUMENTS" > ~/.claude/commands/ai/gpt5.md
 ### Safe Commands (No Security Issues)
 
 ```markdown
-# System prompt editor (file reference only)  
+# System prompt editor (file reference only)
+
 (@)path/to/system/prompt.md
 
 Edit your system prompt above.
@@ -148,6 +152,7 @@ Edit your system prompt above.
 ---
 allowed-tools: Bash(git status:*), Bash(npm list:*)
 ---
+
 Current git status: (!)git status(`)
 Package info: (!)npm list --depth=0(`)
 
@@ -158,6 +163,7 @@ Review project state and suggest next steps.
 
 ```markdown
 # Compare config files
+
 Compare (@)path/to/system.md with (@)project/config.md
 
 Show differences and suggest improvements.
@@ -169,7 +175,6 @@ After creating: `/<command-name> [arguments]`
 
 Example: `/review` or `/ai:gpt5 "explain this code"`
 
-
 ## 🛡 Project Rules (Injected into every command)
 
 1. **NO BROKEN BUILDS:**
@@ -194,7 +199,7 @@ Example: `/review` or `/ai:gpt5 "explain this code"`
 4. **API KEY MANAGEMENT:**
    - Support three loading methods:
      - Server environment variables
-     - NEXT_PUBLIC_ variables (client-side)
+     - NEXT*PUBLIC* variables (client-side)
      - Custom headers (X-OpenAI-Key, X-Anthropic-Key, etc.)
    - Never hardcode API keys
    - Always validate key availability before use
@@ -219,4 +224,4 @@ Example: `/review` or `/ai:gpt5 "explain this code"`
      - Code examples
      - Testing notes
    - Document expected behavior and edge cases
-   - Include real-world output examples
+   - Include real-world output examples

package/.skills/api-create/SKILL.md

@@ -22,6 +22,7 @@ allowed-tools: WebSearch WebFetch mcp__context7 mcp__github AskUserQuestion Read
 **YOU MUST USE THE `AskUserQuestion` TOOL AT EVERY CHECKPOINT.**
 
 This workflow requires REAL user input at each phase. You are **FORBIDDEN** from:
+
 - Self-answering questions
 - Assuming user responses
 - Proceeding without explicit user confirmation
@@ -39,9 +40,9 @@ At every `[Y/n]` or multiple-choice prompt in this workflow, you MUST call the `
       "header": "Phase",
       "multiSelect": false,
       "options": [
-        {"label": "Option A", "description": "What this option means"},
-        {"label": "Option B", "description": "What this option means"},
-        {"label": "Other", "description": "I'll type my own answer"}
+        { "label": "Option A", "description": "What this option means" },
+        { "label": "Option B", "description": "What this option means" },
+        { "label": "Other", "description": "I'll type my own answer" }
       ]
     }
   ]
@@ -49,6 +50,7 @@ At every `[Y/n]` or multiple-choice prompt in this workflow, you MUST call the `
 ```
 
 **CRITICAL REQUIREMENTS:**
+
 - `header`: Max 12 characters (e.g., "Scope", "Research", "Format")
 - `options`: 2-4 options, each with `label` (1-5 words) and `description`
 - `multiSelect`: Required boolean (true for checkboxes, false for radio)
@@ -60,6 +62,7 @@ At every `[Y/n]` or multiple-choice prompt in this workflow, you MUST call the `
 ### Violation Detection
 
 The enforcement hooks will BLOCK your progress if:
+
 - `user_question_asked` is false for any phase
 - `user_confirmed`/`user_approved`/`user_completed` is false
 - `phase_exit_confirmed` is false (user must explicitly approve proceeding to next phase)
@@ -72,27 +75,35 @@ If you see "BLOCKED" messages, it means you skipped user interaction.
 **Every phase requires an EXIT CONFIRMATION question** before proceeding to the next phase. This prevents Claude from self-answering and moving on without explicit user approval.
 
 The exit confirmation question MUST:
+
 1. Summarize what was accomplished in the current phase
 2. Ask if user is ready to proceed to the next phase
 3. Include options like "Yes, proceed", "No, I have changes", "Add more"
 
 Example exit confirmation:
+
 ```json
 {
-  "questions": [{
-    "question": "Phase complete. Research found 5 sources. Ready to proceed to Interview phase?",
-    "header": "Proceed",
-    "multiSelect": false,
-    "options": [
-      {"label": "Yes, proceed", "description": "Move to next phase"},
-      {"label": "No, more research", "description": "I need additional research on [topic]"},
-      {"label": "Review sources", "description": "Show me what was found"}
-    ]
-  }]
+  "questions": [
+    {
+      "question": "Phase complete. Research found 5 sources. Ready to proceed to Interview phase?",
+      "header": "Proceed",
+      "multiSelect": false,
+      "options": [
+        { "label": "Yes, proceed", "description": "Move to next phase" },
+        {
+          "label": "No, more research",
+          "description": "I need additional research on [topic]"
+        },
+        { "label": "Review sources", "description": "Show me what was found" }
+      ]
+    }
+  ]
 }
 ```
 
 The `phase_exit_confirmed` flag is automatically set when:
+
 1. An `AskUserQuestion` is called with a question containing words like "proceed", "continue", "ready", "confirm", "approve"
 2. The user responds with an affirmative answer (yes, proceed, confirm, approve, etc.)
 
@@ -100,6 +111,37 @@ Both conditions must be true for the flag to be set.
 
 ---
 
+## Progress Tracking (TodoWrite Integration)
+
+**FIRST ACTION:** Initialize the todo list to show user real-time progress:
+
+```
+TodoWrite([
+  {"content": "Phase 1: Disambiguation", "status": "in_progress", "activeForm": "Clarifying API terms"},
+  {"content": "Phase 2: Scope", "status": "pending", "activeForm": "Confirming endpoint scope"},
+  {"content": "Phase 3: Initial Research", "status": "pending", "activeForm": "Researching documentation"},
+  {"content": "Phase 4: Interview", "status": "pending", "activeForm": "Gathering requirements"},
+  {"content": "Phase 5: Deep Research", "status": "pending", "activeForm": "Deep diving documentation"},
+  {"content": "Phase 6: Schema", "status": "pending", "activeForm": "Creating Zod schema"},
+  {"content": "Phase 7: Environment", "status": "pending", "activeForm": "Verifying API keys"},
+  {"content": "Phase 8: TDD Red", "status": "pending", "activeForm": "Writing failing tests"},
+  {"content": "Phase 9: TDD Green", "status": "pending", "activeForm": "Implementing to pass tests"},
+  {"content": "Phase 10: Verify", "status": "pending", "activeForm": "Verifying against docs"},
+  {"content": "Phase 11: Code Review (Greptile)", "status": "pending", "activeForm": "Running AI code review"},
+  {"content": "Phase 12: Refactor", "status": "pending", "activeForm": "Fixing issues, cleaning up"},
+  {"content": "Phase 13: Documentation", "status": "pending", "activeForm": "Updating documentation"},
+  {"content": "Phase 14: Completion", "status": "pending", "activeForm": "Final commit and PR"}
+])
+```
+
+**At each phase transition:**
+
+1. Mark current phase as `completed`
+2. Mark next phase as `in_progress`
+3. This gives the user visual progress during long workflows
+
+---
+
 ## Key Principles
 
 1. **Loop Until Green** - Every verification phase loops back if not successful
@@ -107,6 +149,7 @@ Both conditions must be true for the flag to be set.
 3. **Adaptive Research** - Propose searches based on context, not shotgun approach
 4. **Self-Documenting** - State file captures everything for re-grounding
 5. **Verify After Green** - Re-research docs to catch memory-based implementation errors
+6. **Visual Progress** - Update TodoWrite at every phase for user visibility
 
 ## Complete Phase Flow
 
@@ -166,9 +209,34 @@ Both conditions must be true for the flag to be set.
 └───────────────────────────────────────────────────────────┘
         │
         ▼
-┌─ PHASE 3: INITIAL RESEARCH ───────────────────────────────┐
+┌─ PHASE 3: INITIAL RESEARCH (Parallel Subagents) ──────────┐
 │                                                           │
-│ Execute 2-3 initial searches:                             │
+│ Execute 2-3 initial searches using PARALLEL SUBAGENTS:    │
+│                                                           │
+│ ** ASYNC PARALLEL RESEARCH **                             │
+│ For faster research, spawn multiple subagents in parallel │
+│ using the Task tool with different research strategies:   │
+│                                                           │
+│   // Spawn 3 parallel research agents                     │
+│   Task({                                                  │
+│     subagent_type: "parallel-researcher",                 │
+│     prompt: "Research [API] via Context7 - find all       │
+│              endpoints, parameters, webhooks",            │
+│     model: "haiku"                                        │
+│   })                                                      │
+│   // Press Ctrl+B to background this agent                │
+│                                                           │
+│   Task({                                                  │
+│     subagent_type: "parallel-researcher",                 │
+│     prompt: "WebSearch [API] official docs and guides",   │
+│     model: "haiku"                                        │
+│   })                                                      │
+│   // Press Ctrl+B to background this agent                │
+│                                                           │
+│   // Use /tasks to monitor progress of background agents  │
+│                                                           │
+│ ** ALTERNATIVE: Sequential Research **                    │
+│ If not using parallel agents, execute sequentially:       │
 │   • Context7: "[library/api name]"                        │
 │   • WebSearch: "[name] official documentation"            │
 │   • WebSearch: "site:[domain] api reference"              │
@@ -387,9 +455,46 @@ Both conditions must be true for the flag to be set.
 └───────────────────────────────────────────────────────────┘
         │
         ▼
-┌─ PHASE 11: TDD REFACTOR ──────────────────────────────────┐
+┌─ PHASE 11: AI CODE REVIEW (Greptile) ────────────────────┐
+│                                                           │
+│ Run Greptile BEFORE refactoring to catch issues early:    │
+│   • Bug detection with full codebase context              │
+│   • Security vulnerability scanning (OWASP top 10)        │
+│   • Performance issue identification                      │
+│   • Code quality and maintainability checks               │
+│                                                           │
+│ Greptile uses query endpoint to review your diff:         │
+│   POST https://api.greptile.com/v2/query                  │
+│   - Analyzes changes against entire codebase              │
+│   - Returns issues with file:line references              │
+│   - Provides actionable fix suggestions                   │
+│                                                           │
+│ ⚠️ REQUIRED: Use AskUserQuestion tool:                    │
+│                                                           │
+│   AskUserQuestion({                                       │
+│     questions: [{                                         │
+│       question: "Greptile review complete. Score: [N]/10  │
+│                  Issues found: [list]. How to proceed?",  │
+│       header: "Review",                                   │
+│       multiSelect: false,                                 │
+│       options: [                                          │
+│         {"label": "Fix all issues", "description": "Address each issue in Phase 12"},│
+│         {"label": "Fix critical only", "description": "Skip medium/low priority"},│
+│         {"label": "Skip review", "description": "Proceed without fixes (not recommended)"}│
+│       ]                                                   │
+│     }]                                                    │
+│   })                                                      │
+│                                                           │
+│ WAIT for user response. Do NOT auto-skip issues.          │
+│ REQUIRES: GREPTILE_API_KEY + GITHUB_TOKEN in .env         │
+│ ──── Loop back to Phase 9 if major issues found ────      │
+└───────────────────────────────────────────────────────────┘
+        │
+        ▼
+┌─ PHASE 12: TDD REFACTOR ──────────────────────────────────┐
 │                                                           │
-│ Clean up code while tests stay green:                     │
+│ Fix Greptile issues AND clean up code:                    │
+│   • Fix all issues from Phase 11 code review              │
 │   • Extract reusable patterns                             │
 │   • Improve error messages                                │
 │   • Add JSDoc comments                                    │
@@ -399,7 +504,7 @@ Both conditions must be true for the flag to be set.
 └───────────────────────────────────────────────────────────┘
         │
         ▼
-┌─ PHASE 12: DOCUMENTATION ─────────────────────────────────┐
+┌─ PHASE 13: DOCUMENTATION ─────────────────────────────────┐
 │                                                           │
 │ Update documentation files, then:                         │
 │                                                           │
@@ -410,10 +515,11 @@ Both conditions must be true for the flag to be set.
 │       question: "Documentation checklist: [list files].   │
 │                  All documentation complete?",            │
 │       header: "Docs",                                     │
+│       multiSelect: false,                                 │
 │       options: [                                          │
-│         "Yes, all documentation is done",                 │
-│         "No, I need to add something (I'll describe)",    │
-│         "Skip docs for now (not recommended)"             │
+│         {"label": "Yes, complete", "description": "All docs updated"},│
+│         {"label": "Need to add more", "description": "I'll describe what's missing"},│
+│         {"label": "Skip for now", "description": "Not recommended"}│
 │       ]                                                   │
 │     }]                                                    │
 │   })                                                      │
@@ -424,16 +530,18 @@ Both conditions must be true for the flag to be set.
 └───────────────────────────────────────────────────────────┘
         │
         ▼
-┌─ PHASE 13: COMPLETION ────────────────────────────────────┐
+┌─ PHASE 14: COMPLETION ────────────────────────────────────┐
 │                                                           │
 │ Final verification:                                       │
 │   • All tests passing                                     │
 │   • 100% coverage                                         │
 │   • TypeScript compiles                                   │
+│   • Code review issues addressed                          │
 │   • Docs updated                                          │
 │   • State file shows all phases complete                  │
 │                                                           │
 │ Run /commit to create semantic commit.                    │
+│ Run /pr to create pull request.                           │
 └───────────────────────────────────────────────────────────┘
 ```
 
@@ -456,8 +564,10 @@ All phases are tracked in `.claude/api-dev-state.json`:
     "tdd_red": { "status": "complete", "test_count": 23 },
     "tdd_green": { "status": "complete" },
     "verify": { "status": "complete", "gaps_found": 2, "gaps_fixed": 2 },
+    "code_review": { "status": "complete", "score": 9, "issues_found": 1, "issues_fixed": 1 },
     "tdd_refactor": { "status": "complete" },
-    "documentation": { "status": "complete" }
+    "documentation": { "status": "complete" },
+    "completion": { "status": "complete" }
   }
 }
 ```
@@ -477,16 +587,17 @@ This command creates:
 
 ## Hooks That Enforce This Workflow
 
-| Phase | Hook | Purpose |
-|-------|------|---------|
-| 0 | `enforce-external-research.py` | Detects API terms, requires disambiguation |
-| 2-4 | `track-tool-use.py` | Logs all research, tracks turns |
-| 7-8 | `enforce-research.py` | Blocks Write if no research done |
-| 7-8 | `enforce-interview.py` | Injects interview decisions |
-| 8 | `verify-implementation.py` | Blocks route if no test file |
-| 9 | `verify-after-green.py` | Triggers verification after tests pass |
-| All | `periodic-reground.py` | Re-grounds every 7 turns |
-| 11 | `api-workflow-check.py` | Blocks completion if docs incomplete |
+| Phase | Hook                           | Purpose                                    |
+| ----- | ------------------------------ | ------------------------------------------ |
+| 0     | `enforce-external-research.py` | Detects API terms, requires disambiguation |
+| 2-4   | `track-tool-use.py`            | Logs all research, tracks turns            |
+| 7-8   | `enforce-research.py`          | Blocks Write if no research done           |
+| 7-8   | `enforce-interview.py`         | Injects interview decisions                |
+| 8     | `verify-implementation.py`     | Blocks route if no test file               |
+| 9     | `verify-after-green.py`        | Triggers verification after tests pass     |
+| 10    | `run-code-review.py`           | Triggers Greptile AI code review           |
+| All   | `periodic-reground.py`         | Re-grounds every 7 turns                   |
+| 14    | `api-workflow-check.py`        | Blocks completion if docs incomplete       |
 
 <claude-commands-template>
 ## Project-Specific Rules
@@ -497,7 +608,7 @@ This command creates:
 4. **AI SDK**: Use Vercel AI SDK 5.0.11 patterns from `/src/v2/docs/ai-sdk-catalog.json`
 5. **Package Manager**: Use `pnpm` for all operations
 6. **Documentation**: Follow patterns in `/src/v2/docs/Main Doc – V2 Development Patterns.md`
-7. **API Keys**: Support three methods (env, NEXT_PUBLIC_, custom headers)
+7. **API Keys**: Support three methods (env, NEXT*PUBLIC*, custom headers)
 8. **Test Command**: `pnpm test:run` before commits
 
 ## Never Skip
@@ -509,4 +620,4 @@ This command creates:
 - Phase 10 (Verify) - Re-research after Green
 - Phase 12 (Documentation) - Keep docs in sync
 - Coverage verification - 100% required
-</claude-commands-template>
+  </claude-commands-template>

package/.skills/api-env/SKILL.md

@@ -55,6 +55,7 @@ Status: BLOCKED - Cannot proceed without FIRECRAWL_API_KEY
 ## API Key Support
 
 The project supports three methods:
+
 1. Server env: `OPENAI_API_KEY=sk-...`
 2. Client env: `NEXT_PUBLIC_OPENAI_API_KEY=sk-...`
 3. Custom headers: `X-OpenAI-Key: sk-...`