@hustle-together/api-dev-tools 3.12.3 → 4.5.1

package/CHANGELOG.md

@@ -2,6 +2,581 @@
 
 All notable changes to this project will be documented in this file.
 
+## [4.5.0] - 2025-12-30
+
+### Added - Implementation Gap Fixes (10 Features Now Actually Work)
+
+This release fixes 10 documented features that were not actually implemented.
+
+#### Workflow Logging (All Sessions)
+
+- **`hooks/hook_utils.py`** - Shared logging utility for all hooks
+  - `log_workflow_event()` - Log any workflow event with timestamp
+  - Event types: session_start, phase_transition, interview_decision, auto_answer, promise_emitted
+  - Logs stored in `.claude/workflow-logs/{workflow_id}.json`
+- Logging added to: `session-startup.py`, `auto-answer.py`, `enforce-interview.py`, `completion-promise-detector.py`
+
+#### Directory & Registry Creation
+
+- **Session startup now creates required directories:**
+  - `.claude/workflow-logs/` - Audit trail for all sessions
+  - `.claude/adrs/` - Architecture Decision Records
+  - `.claude/adr-requests/` - Pending ADR research requests
+  - `.claude/research/` - Cached research data
+- **Registry auto-creation:** `.claude/registry.json` created from template on first session
+
+#### `--dry-run` Flag (NEW)
+
+- **`hooks/enforce-dry-run.py`** - Blocks Write/Edit when dry-run active
+  - Allows full workflow preview without file modifications
+  - Shows what WOULD be written at each step
+  - Useful for testing autonomous mode safely
+
+#### `--max-iterations` Enforcement (NEW)
+
+- **Per-phase iteration limits** now enforced
+  - Default: 25 iterations per phase (from `hustle-build-defaults.json`)
+  - Blocks with clear message when limit exceeded
+  - Prevents infinite loops in autonomous mode
+- Added to `completion-promise-detector.py`
+
+#### `--resume` Flag Support (NEW)
+
+- **`hook_utils.py`** - Resume functionality
+  - `handle_resume(workflow_id)` - Restore workflow from logs
+  - `list_resumable_workflows()` - Show all resumable workflows
+  - `snapshot_state_to_log()` - Save state for future resume
+
+#### Ralph Wiggum Skills (NEW)
+
+- **`.skills/ralph-status/SKILL.md`** - Show loop status
+  - Current phase and iteration
+  - Active promises
+  - Elapsed time
+- **`.skills/ralph-continue/SKILL.md`** - Resume paused loops
+  - Clear active promise
+  - Reset iteration counters
+  - Continue from last phase
+
+#### Parallel Execution (NEW)
+
+- **`hooks/parallel-orchestrator.py`** - Git worktree coordination
+  - Creates isolated worktrees for parallel workflows
+  - Injects shared decisions to avoid re-interviewing
+  - Merges results when complete
+- **`.skills/parallel-spawn/SKILL.md`** - Spawn parallel agents
+  - `/parallel-spawn api:users api:products api:orders`
+  - Creates worktrees, spawns background Task agents
+  - Monitor with `/parallel-status`, merge with `/parallel-merge`
+
+### Changed
+
+- **`hooks/session-startup.py`** v4.5.0
+  - Now calls `ensure_directories()` and `ensure_registry()`
+  - Logs `session_start` event
+  - Initializes flags structure for dry-run/resume/parallel
+
+- **`hooks/completion-promise-detector.py`** v4.5.0
+  - Now logs `promise_emitted` events
+  - Enforces `--max-iterations` limit
+  - Logs `iteration_limit_exceeded` when hit
+
+- **`templates/settings.json`** - New hooks registered
+  - `enforce-dry-run.py` in PreToolUse (Write|Edit)
+  - `parallel-orchestrator.py` in SessionStart and UserPromptSubmit
+  - `completion-promise-detector.py` in PostToolUse (Bash, Write|Edit)
+
+### Architecture
+
+```
+┌─────────────────────────────────────────────────────────────────┐
+│ v4.5.0 - IMPLEMENTATION GAPS FIXED                              │
+├─────────────────────────────────────────────────────────────────┤
+│                                                                 │
+│  NOW WORKING:                                                   │
+│  ✓ Workflow logging (all sessions)                              │
+│  ✓ Directory creation (.claude/workflow-logs/, .claude/adrs/)   │
+│  ✓ Registry auto-creation from template                         │
+│  ✓ --dry-run flag (preview without writes)                      │
+│  ✓ --max-iterations enforcement (per-phase limits)              │
+│  ✓ --resume flag (restore from workflow logs)                   │
+│  ✓ /ralph-status skill                                          │
+│  ✓ /ralph-continue skill                                        │
+│  ✓ --parallel flag (git worktree coordination)                  │
+│  ✓ /parallel-spawn skill (concurrent agents)                    │
+│                                                                 │
+│  FILES ADDED:                                                   │
+│  + hooks/enforce-dry-run.py                                     │
+│  + hooks/parallel-orchestrator.py                               │
+│  + .skills/ralph-status/SKILL.md                                │
+│  + .skills/ralph-continue/SKILL.md                              │
+│  + .skills/parallel-spawn/SKILL.md                              │
+│                                                                 │
+│  FILES MODIFIED:                                                │
+│  ~ hooks/hook_utils.py (expanded with logging, resume, etc.)    │
+│  ~ hooks/session-startup.py (directory/registry creation)       │
+│  ~ hooks/auto-answer.py (uses shared logging)                   │
+│  ~ hooks/enforce-interview.py (logs decisions)                  │
+│  ~ hooks/completion-promise-detector.py (iterations + logging)  │
+│  ~ templates/settings.json (new hooks registered)               │
+│                                                                 │
+└─────────────────────────────────────────────────────────────────┘
+```
+
+---
+
+## [4.4.0] - 2025-12-30
+
+### Added - Deep ADR Research (Substantive Trade-offs)
+
+**Problem solved:** ADRs were previously created with placeholder content (`_To be filled based on research_`). Now ADRs contain real, researched pros/cons.
+
+- **`.claude/agents/adr-researcher.md`** - Parallel research agent (Haiku)
+  - Researches a single technology option
+  - Fetches official documentation
+  - Extracts: pros, cons, pricing, best-for, limitations
+  - Returns structured JSON for ADR creation
+
+- **`.skills/adr-deep-research/SKILL.md`** - Deep research skill
+  - Reads pending requests from `.claude/adr-requests/`
+  - Spawns parallel `adr-researcher` agents (one per option)
+  - Merges results into substantive ADR document
+  - Updates registry with ADR metadata
+
+- **`.claude/adr-requests/`** - Research request directory
+  - `pending-{category}.json` - Awaiting research
+  - Processed requests archived after ADR creation
+
+### Changed - ADR Flow (Request-Based)
+
+- **`hooks/generate-adr-options.py`** v2.0 - Creates requests, not placeholders
+  - Previously: Created ADR with empty pros/cons
+  - Now: Creates research REQUEST file
+  - Injects context: "Run `/adr-deep-research {category}`"
+  - AI then runs deep research before interview
+
+- **`.skills/api-research/SKILL.md`** - ADR integration documented
+  - Explains automatic ADR detection during research
+  - Documents the Research → Deep Research → Interview flow
+  - Links to `/adr-deep-research` skill
+
+### Architecture
+
+```
+┌─────────────────────────────────────────────────────────────────┐
+│ DEEP ADR RESEARCH FLOW                                          │
+├─────────────────────────────────────────────────────────────────┤
+│                                                                 │
+│  1. /api-research runs                                          │
+│     ↓                                                           │
+│  2. generate-adr-options.py detects decision                    │
+│     → Creates: .claude/adr-requests/pending-database.json       │
+│     → Injects: "Run /adr-deep-research database"                │
+│     ↓                                                           │
+│  3. /adr-deep-research database                                 │
+│     → Spawns 3 parallel adr-researcher agents                   │
+│     → Each fetches official docs, extracts pros/cons            │
+│     ↓                                                           │
+│  4. Creates: .claude/adrs/ADR-0001-database-choice.md           │
+│     → Real pros/cons from documentation                         │
+│     → Pricing, limitations, best-for recommendations            │
+│     → Source URLs for verification                              │
+│     ↓                                                           │
+│  5. Interview references ADR for informed decision              │
+│                                                                 │
+└─────────────────────────────────────────────────────────────────┘
+```
+
+---
+
+## [4.3.2] - 2025-12-30
+
+### Added - Master Phase Reference
+
+- **`docs/PHASE_REFERENCE.md`** - Complete 14-phase audit reference
+  - Phase Status Matrix with hooks, skills, docs, implementation status
+  - All 14 phases documented with purpose, requirements, exit conditions
+  - Feature Implementation Matrix (ADR, Auto-Answer, Ralph Wiggum, etc.)
+  - All Slash Commands organized by category
+  - Configuration files reference
+
+### Added - ADR Integration (Research → ADR → Interview)
+
+- **`hooks/generate-adr-options.py`** - Creates ADRs during research phase
+  - Detects multiple options for significant decisions (database, auth, cache, etc.)
+  - Creates `.claude/adrs/NNNN-category-choice.md` with options and trade-offs
+  - Configurable decision categories in `hustle-build-defaults.json`
+- **`hooks/update-adr-decision.py`** - Updates ADRs after interview
+  - Changes status from PROPOSED to ACCEPTED
+  - Records user's decision and reasoning
+- **Registry `adrs` section** - Track all ADRs with metadata
+  - Status, category, decision, options considered
+  - Links to endpoint and ADR file
+
+### Fixed - Auto-Answer Now Actually Works
+
+- **`hooks/auto-answer.py`** - Uses `continue: false` pattern
+  - Previously: Returned `continue: true` with context (didn't skip question UI)
+  - Now: Returns `continue: false` with answer in reason (AI uses the answer)
+  - Shows brief "Auto-Selected: X" message before proceeding
+  - Logs all auto-answers to `.claude/workflow-logs/`
+
+### Changed
+
+- **Skills updated with `(Recommended)` options** - Key skills updated
+  - `api-create/SKILL.md` - All 14 phase options marked with (Recommended)
+  - `hustle-build/SKILL.md` - Decomposition and interview options
+  - `hustle-brand/SKILL.md` - Interview flow and voice options
+  - Enables auto-answer hook to detect recommended options
+- **`docs/CONFIGURATION.md`** - Added Auto-Answer Selection documentation
+  - Explains 3-tier priority: defaults → (Recommended) → keyword fallback
+  - Documents scoring algorithm for comprehensive selection
+- **README.md** - Comprehensive documentation update
+  - All Slash Commands section with 9 categories (40+ commands)
+  - Phase-specific commands with phase numbers
+  - Configuration section with all config files
+  - Updated hook count (54) and component counts
+  - Links to new PHASE_REFERENCE.md as master reference
+  - New FAQ entries for ADRs and autonomous mode
+- **`templates/hustle-build-defaults.json`** v3.1.0
+  - Added `adr` section with configurable `significant_decisions`
+  - Added `min_options_for_adr` threshold (default: 2)
+- **`templates/registry.json`** v1.5.0
+  - Added `adrs` section for tracking Architecture Decision Records
+- **`templates/settings.json`** - Registered new ADR hooks
+
+---
+
+## [4.3.1] - 2025-12-30
+
+### Added - Configuration Documentation
+
+- **`docs/CONFIGURATION.md`** - Comprehensive configuration guide
+  - All configurable options in `hustle-build-defaults.json`
+  - How autonomous/skip interview mode works
+  - Logging locations for post-hoc review
+  - Question mappings and customization examples
+
+### Added - Architecture Decision Records (ADRs)
+
+- **`docs/ARCHITECTURE_DECISION_RECORDS.md`** - ADR integration guide
+  - When to create ADRs (Interview & Deep Research phases)
+  - ADR template with context, decision drivers, consequences
+  - Storage in `.claude/adrs/` with index.json registry
+  - Integration with dashboard
+- **`templates/adr-viewer/_components/ADRViewer.tsx`** - Dashboard component
+  - Filter by status (proposed, accepted, deprecated, superseded)
+  - Filter by phase (initial research, interview, deep research)
+  - Search across ADR titles and endpoints
+  - Detail view with markdown rendering
+
+### Changed - Autonomous Mode Now Default
+
+- **`hustle-build-defaults.json` v3.0.0** - Autonomous mode ON by default
+  - `autonomous.enabled: true` - No `--auto` flag needed
+  - `autonomous.skip_interviews: true` - Uses comprehensive defaults
+  - `autonomous.ralph_wiggum_loops: true` - Iterative phases loop automatically
+  - `autonomous.auto_fix_visual_issues: true` - Visual QA fixes without prompting
+  - `autonomous.auto_fix_review_issues: true` - Code review fixes without prompting
+- **`hooks/auto-answer.py`** - Now checks defaults file for autonomous mode
+  - Falls back to template defaults if no project-specific file
+  - No state file required for auto-answering to work
+- **`docs/AUTONOMOUS_LOOPS.md`** - Updated to reflect default-on behavior
+
+### How to Disable
+
+Set `autonomous.enabled: false` in `.claude/hustle-build-defaults.json` to require manual interviews.
+
+---
+
+## [4.3.0] - 2025-12-29
+
+### Added - Parallel Autonomous Workflow
+
+- **Up to 5 Opus agents** running in parallel across git worktrees
+- **`/hustle-build --parallel`** flag for parallel execution
+- **`docs/PARALLEL_AUTONOMOUS_WORKFLOW.md`** - Complete architecture documentation
+- **Shared context injection** - Orchestrator decisions passed to all agents
+- **Merge coordinator** - Combines registry entries, resolves conflicts
+- **Agent status monitoring** - Track progress across all parallel agents
+
+### Added - Multi-Pass Code Review
+
+- **4-pass review system** with deterministic checklists:
+  - Pass 1: Logic & Bugs (null handling, off-by-one, race conditions)
+  - Pass 2: Security (auth, authorization, injection, data exposure)
+  - Pass 3: Performance (N+1, memory leaks, re-renders)
+  - Pass 4: Miscellaneous (AI judgment - clarity, patterns, docs)
+- **Per-item pass/fail tracking** - Every checklist item answered
+- **Combined summary report** - Aggregates all pass findings
+- **Review Dashboard Template** (`templates/review-dashboard/page.tsx`) - Visual display of multi-pass results
+
+### Added - Max Iterations Flag
+
+- **`/hustle-build --max-iterations [N]`** - Prevent infinite loops
+- **Per-phase limits** - Different defaults for each phase type
+- **Graceful degradation** - Creates partial PR when limit reached
+- **Session archival** - Logs interrupted workflows for review
+
+### Added - Documentation System
+
+- **`/docs-update`** skill - Ensures README/CHANGELOG stay current
+- **`hooks/docs-update-check.py`** - PostToolUse hook reminds about doc updates
+- **`docs/BRAND_GUIDE.md`** - Complete brand system setup documentation
+- **README.md** updated to v4.3.0 with new counts (38+ skills, 24 hooks, 6 templates)
+
+### Added - Research Enhancements
+
+- **TOC Scraping** - Fetches documentation table of contents before interview
+- **Comprehensive Discovery Checklist** - Auth, endpoints, params, webhooks, SDKs, errors
+- **Enhanced api-research skill** with discovery flow diagram
+
+### Added - Test Mode Enhancements
+
+- **`/hustle-build --auto`** - Full autonomous builds without prompts
+- **`templates/hustle-build-defaults.json`** v2.0.0 - Configurable default answers
+- **Per-category defaults** - orchestrator, api, component, page, combined, testing
+- **Use cases table** - CI/CD integration, demo mode, quick testing
+
+### Enhanced - ShadCN Skill
+
+- **Design System Architecture** diagram showing brand → CSS → Tailwind → ShadCN flow
+- **CSS Variable Mapping** table (Brand Guide → CSS Variable → ShadCN Usage)
+- **Complete globals.css template** with light/dark mode support
+- **`/shadcn sync`** command for brand guide synchronization
+- **Registry Integration** section showing tracked values
+
+### Added - Final Gap Closures
+
+- **Completion Promise Detection (Ralph Wiggum)** - Autonomous loop self-termination
+  - `hooks/completion-promise-detector.py` - Detects `<promise>DONE</promise>` signals
+  - `.skills/ralph-loop/SKILL.md` - `/ralph-loop` skill for autonomous tasks
+  - `/ralph-status` and `/ralph-continue` commands
+  - **Iterative phase integration:**
+    - `/test-review` outputs `<promise>REVIEW_CLEAN</promise>` when all passes clean
+    - `/refactor` outputs `<promise>REFACTORED</promise>` when complete
+    - `/test-visual` outputs `<promise>VISUAL_CLEAN</promise>` when all viewports pass
+  - `docs/AUTONOMOUS_LOOPS.md` - Complete pattern documentation with Geoffrey Huntley credit
+- **Schema Lint ESLint Plugin** - Zod best practices enforcement
+  - `templates/eslint-plugin-zod-schema/` - Full ESLint plugin
+  - Rules: require-description, consistent-naming, no-unsafe-defaults, prefer-strict
+  - Recommended and strict configs
+- **Dependency Audit Workflow** - Security scanning in CI
+  - `templates/github-workflows/security.yml` - Complete GitHub Actions workflow
+  - Dependency audit, license check, secret scan, SAST
+  - Auto-detects npm/pnpm/yarn
+- **Credits & Acknowledgments** section added to README
+  - Geoffrey Huntley (Ralph Wiggum Pattern)
+  - Kent Beck (TDD Workflow)
+  - Context7 and GitHub MCPs
+
+### Changed
+
+- **`/test-review`** - Now uses multi-pass system with promise completion signal
+- **`/refactor`** - Adds promise completion signal for autonomous loops
+- **`/test-visual`** - Adds promise completion signal for visual QA loops
+- **`/hustle-build`** - Supports parallel, max-iterations, and auto flags
+- **ROADMAP.md** - Coverage increased from 99% to 100% (MCP limit skipped by design)
+
+---
+
+## [4.2.0] - 2025-12-29
+
+### Added - Brand Guide System
+
+- **`/hustle-brand`** - Comprehensive brand guide creator with interview-driven discovery
+  - Visual identity (colors, typography, spacing)
+  - Motion & animation preferences (GSAP, Framer Motion, CSS)
+  - Voice & tone guidelines (professional, friendly, technical, playful)
+  - Custom elements (terminal animations, gradients, Three.js, etc.)
+  - Do's and Don'ts for consistency
+- **Brand Page Template** (`templates/brand-page/page.tsx`) - Living showcase with:
+  - Color palette with copy-to-clipboard
+  - Typography scale demonstration
+  - Button states (all variants, sizes, loading)
+  - Form elements showcase
+  - Animation examples
+  - Voice examples based on tone
+- **Registry brand_guide section** (v1.4.0) - Tracking for:
+  - Interview completion status
+  - Custom elements selected
+  - Voice configuration
+  - ShadCN integration status
+
+### Added - ShadCN Integration
+
+- **`/shadcn`** - ShadCN documentation skill with 15-day freshness auto-update
+  - `docs [component]` - Get latest component docs
+  - `add [components]` - Install with brand theme
+  - `status` - Show installed components and freshness
+  - Auto-refresh via Context7 when >15 days old
+
+### Changed - Visual Testing Output
+
+- **Enhanced Haiku analysis output** - Now includes:
+  - Haiku's detailed reasoning for each finding
+  - Screenshot file paths (clickable links)
+  - Storybook URLs for each viewport
+  - Category-by-category pass/fail breakdown
+
+---
+
+## [4.1.0] - 2025-12-29
+
+### Coverage: 77% → 91%
+
+Major release completing all planned test skills, visual testing, token tracking, and infrastructure improvements.
+
+### Added - Test Skills (8 new skills)
+
+- **`/test-unit`** - Run Vitest unit tests with coverage thresholds
+- **`/test-e2e`** - Run Playwright E2E with cross-browser reporting
+- **`/test-visual`** - Storybook visual + interaction tests with 7 viewports
+- **`/test-all`** - Complete suite: unit → e2e → visual → builds → review
+- **`/test-review`** - **Tiered security strategy**: ESLint 100% + AI on critical paths
+- **`/test-builds`** - **Browser-only testing** (Chrome/Firefox/WebKit = all platforms)
+- **`/test-debug`** - DOM snapshots, root cause analysis
+- **`/visual-qa`** - Full visual QA with Haiku analysis
+
+### Added - Visual Testing Enhancements
+
+- **7 Viewports**: Mobile portrait, notch, landscape; tablet portrait/landscape; small/large desktop
+- **Safe Area Insets**: iOS notch support (top: 47px, bottom: 34px)
+- **Haiku Visual Analyzer Agent**: AI-powered screenshot analysis
+- **Enhanced Output**: Haiku's detailed reasoning for each issue
+- **Quick Links**: Screenshot paths, Storybook URLs, Playwright reports
+
+### Added - Registry Expansion (v1.3.0)
+
+- **`routes`** - API routes + page routes tracking
+- **`env_vars`** - Required environment variables with docs links
+- **`services`** - External dependencies (Stripe, Supabase, OpenAI, etc.)
+- **`webhooks`** - Incoming webhook endpoints with signature config
+
+### Added - State Management (v3.11.0)
+
+- **Session Archives** - Completed/interrupted workflow history
+- **Learnings Aggregation** - Cross-session pattern learning
+- **Re-grounding Integration** - Full infrastructure awareness in 7-turn reminders
+
+### Added - Security
+
+- **Tiered Security Review**: ESLint on ALL files + AI on changed/critical paths
+- **AI Security Patterns**: SQL injection, auth bypass, CSRF, IDOR, mass assignment, data exposure
+- **Security Deny Rules**: In `templates/settings.json`
+
+### Added - Token Tracking
+
+- **`/token-report`** skill with ccusage integration
+- **Per-phase timestamps** in state for cost correlation
+- **Context capacity warning** hook at 80% usage
+
+### Added - Documentation
+
+- **`docs/REGROUNDING.md`** v4.1.0 - Registry integration section
+- **`docs/PRE-COMMIT-SETUP.md`** - lint-staged configuration
+- **`docs/ESLINT-CONFIG.md`** - Type-aware rules setup
+
+### Changed
+
+- **`/test-builds`** - Now browser-only (Chromium/Firefox/WebKit covers Tauri/Capacitor/Electron)
+- **`/test-review`** - Tiered approach for large codebases (ESLint 100% + AI on critical)
+- **`periodic-reground.py`** - Now includes routes, services, webhooks, env vars
+
+### Fixed
+
+- **Source Repository Detection** - Hooks no longer self-enforce on api-dev-tools source
+- **Hook Utils** - New `hooks/hook_utils.py` with `is_source_repository()` function
+
+---
+
+## [3.12.12] - 2025-12-28
+
+### Added
+
+- **Dashboard Page Templates** - All dashboard links now work (no more 404s)
+  - `/docs` - TypeDoc documentation viewer with generation instructions
+  - `/test-results` - Vitest results page with test commands reference
+  - `/playwright-report` - E2E test report viewer with Playwright commands
+
+- **Interactive Query Parameter Builder** - Enhanced API testing UX
+  - Checkbox toggle to include/exclude each parameter in query string
+  - Type-aware inline editors:
+    - Dropdown `<select>` for enum types (order_by, color, orientation)
+    - Number inputs with min/max validation for numeric types
+    - Text inputs with placeholder examples for strings
+  - Auto-updates query string as you check params and edit values
+  - Required params locked on (can't uncheck)
+  - Example preset buttons still work as quick templates
+
+### Changed
+
+- **ParameterDocs component** renamed to `InteractiveParamBuilder` for GET requests
+- Query params now sync bidirectionally between builder and input field
+
+### Documentation
+
+- **[docs/HOOKS.md](./docs/HOOKS.md)** - Complete hook reference (45+ hooks)
+- **[docs/SKILLS.md](./docs/SKILLS.md)** - All slash commands with usage examples
+- **[docs/AGENTS.md](./docs/AGENTS.md)** - Specialized subagent reference
+- **[docs/PLUGIN_ARCHITECTURE.md](./docs/PLUGIN_ARCHITECTURE.md)** - How the plugin system works
+
+---
+
+## [3.12.11] - 2025-12-28
+
+### Added
+
+- **Hustle Dev Dashboard** (`/hustle-dev-dashboard`) - Central hub page template
+  - Links to API Showcase, UI Showcase, Storybook, TypeDoc, Test Results
+  - Shows registry stats (APIs, Combined, Components, Pages)
+  - Quick Commands section with common pnpm commands
+  - Brand-themed design matching Hustle Together style
+
+- **End-of-Workflow Summaries** - All 4 main commands now show completion summaries
+  - `/hustle-api-create` - Shows API quick links and next steps
+  - `/hustle-combine` - Shows combined API info and dashboard link
+  - `/hustle-ui-create` - Shows component preview links and Storybook
+  - `/hustle-ui-create-page` - Shows page URL and E2E test commands
+
+### Changed
+
+- **API Modal UX Improvements** - Better testing experience
+  - Modal widened from `max-w-5xl` to `max-w-7xl` for more content space
+  - "Send Request" button moved to footer (alongside View Tests/Copy Schema)
+  - Reduces scrolling when testing APIs
+
+---
+
+## [3.12.10] - 2025-12-28
+
+### Added
+
+- **TypeDoc Integration** - Generate API documentation from TSDoc comments
+  - `pnpm typedoc` script to generate Markdown docs
+  - `pnpm typedoc:watch` for development mode
+  - `templates/typedoc.json` configuration template
+  - Runs during Phase 13 (Documentation)
+
+- **API Showcase Example Requests** - Pre-built, runnable examples in the testing UI
+  - Example selector UI with clickable buttons
+  - Auto-fills query parameters when clicked
+  - "Copy curl" button for each example
+  - Examples auto-generated from Zod schema parameters
+
+- **Enhanced extract-schema-docs.cjs** - Generates working examples from schemas
+  - Detects required params, enums, and defaults
+  - Builds query strings and curl commands
+  - Outputs `examples` section for registry.json
+
+### Fixed
+
+- **APIModal 404 errors** - Fixed endpoint path building for action-based APIs
+  - Action-based APIs now use query params (`/api/v2/unsplash?action=search`)
+  - No longer incorrectly builds sub-paths (`/api/v2/unsplash/search`)
+
+---
+
 ## [1.0.1] - 2025-12-28
 
 ### Fixed