Update the feature workflow commands to create GitHub issues per phase
instead of per individual task. This reduces issue clutter while still
maintaining granular task tracking via checklists within phase issues.
Key changes:
**publish-to-github.md:**
- Create phase issues instead of individual task issues
- Each phase issue contains the full task checklist from implementation plan
- Add support for `[complex]` marker to break out complex tasks as separate issues
- Update Epic to link to phase issues instead of tasks
- Update github.md template to show phases and optional complex task issues
**create-feature.md:**
- Add documentation for `[complex]` task marker
- Update implementation plan format example with nested sub-tasks
- Add "When to Use [complex]" guidance section
**continue-feature.md:**
- Rewrite workflow to work with phase issues instead of task issues
- Add logic to identify current phase and find unchecked tasks
- Support both phase issue tasks and complex task issues
- Add step to update phase issue checklist after completing tasks
- Update completion reporting for phase-based progress
- Renumber steps (8.x → 7.x) for consistency
**Hybrid approach:**
- Default: one issue per phase with task checklists
- Optional: break out tasks marked `[complex]` or with nested sub-tasks
- Complex task issues link back to parent phase issue
This change results in ~5 issues (1 epic + 4 phases) instead of ~47 issues
for a typical feature, significantly reducing GitHub issue noise.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Extends the /create-feature command to generate a third file that
extracts manual/human tasks into a dedicated checklist for better
visibility.
Changes:
- Add step 4 to create action-required.md with manual tasks
- Add template format with Before/During/After Implementation sections
- Add "no manual steps" template variant
- Update Next Steps to prioritize reviewing action-required.md
- Add note listing common manual task types (API keys, env vars, etc.)
- Bump create-agentic-app version to 1.1.27
Manual tasks (like account creation, environment variables, OAuth setup)
are now surfaced in a separate file so they're harder to miss during
feature implementation.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Add comprehensive documentation for the custom Claude Code slash commands
that enable GitHub-integrated feature development workflow:
- /create-feature: Create feature specs with requirements and implementation plan
- /publish-to-github: Publish features to GitHub Issues and Projects
- /continue-feature: Implement tasks one at a time with automatic GitHub updates
- /checkpoint: Create detailed checkpoint commits
Documentation includes:
- Prerequisites (GitHub CLI setup, project scopes)
- Step-by-step typical workflow guide
- Example terminal session
- Offline mode instructions
- Command files location
Also bumps package version to 1.1.26 for this documentation update.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
This commit introduces a comprehensive feature management workflow that
integrates with GitHub Issues and Projects for better task tracking and
team collaboration.
## New Commands Added
- `/publish-to-github`: Publishes features from /specs to GitHub by:
- Creating an Epic issue with full requirements
- Creating individual task issues for each implementation step
- Setting up a GitHub Project board linked to the repository
- Creating labels for organization (epic, feature/*, phase-*)
- Generating a github.md reference file in the specs folder
- `/continue-feature`: Implements the next available task by:
- Querying open issues for the feature
- Checking task dependencies to find unblocked work
- Updating GitHub Project board status (In Progress -> Done)
- Adding implementation details as issue comments
- Providing fallback to implementation-plan.md when offline
## Updated Commands
- `/create-feature`: Enhanced with clearer structure including:
- Detailed implementation plan format template
- Requirements for atomic, agent-implementable tasks
- Guidance on next steps after feature creation
- Better documentation for the /specs folder structure
## Package Updates
- Bumped version from 1.1.24 to 1.1.25
All changes are mirrored in both the root .claude/commands/ folder
and the create-agentic-app/template/.claude/commands/ folder to ensure
new projects created with the CLI have access to these workflows.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Changes:
- Add Vercel Blob storage hostname pattern (*.public.blob.vercel-storage.com)
to Next.js image domains config for both main project and template
- Bump create-agentic-app version from 1.1.23 to 1.1.24
- Minor whitespace cleanup in storage.ts files
This allows images stored in Vercel Blob to be properly optimized
by Next.js Image component in production deployments.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Remove platform-specific Windows path from additionalDirectories in Claude settings files to ensure cross-platform compatibility across Windows, macOS, and Linux.
Renamed the ui-developer agent to ui-expert for better clarity and
consistency with naming conventions. This change affects both the root
project and the create-agentic-app template.
Changes made:
- Renamed .claude/agents/ui-developer.md to ui-expert.md in both root and template
- Updated agent name and references within agent configuration
- Bumped create-agentic-app version from 1.1.14 to 1.1.15
- Updated corresponding package-lock.json with new version
The agent's functionality and responsibilities remain unchanged - this is
purely a naming improvement to better reflect its role as a UI expert
rather than just a developer.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Successfully upgraded the project from Next.js 15 to Next.js 16 stable release
with all related dependencies and configuration updates.
## Package Upgrades
- Next.js: 15.4.6 → 16.0.3
- React: 19.1.0 → 19.2.0
- React DOM: 19.1.0 → 19.2.0
- @types/react: 19.2.2 → 19.2.5
- @types/react-dom: 19.2.2 → 19.2.3
- eslint-config-next: 15.4.6 → 16.0.3
## Configuration Changes
- Removed --turbopack flag from dev script (now default in Next.js 16)
- Updated lint script from 'next lint' to 'eslint .' for better compatibility
- Migrated ESLint config from FlatCompat to native flat config format
- Removed @eslint/eslintrc dependency (no longer needed)
- Added pnpm overrides for React type definitions
## ESLint Configuration
- Converted from FlatCompat compatibility layer to direct import
- Now using eslint-config-next/core-web-vitals directly
- Added ignores for .next, node_modules, .cache, dist, and build directories
- Simplified configuration structure for better maintainability
## TypeScript Configuration
- Auto-updated by Next.js 16 build process
- Changed jsx from 'preserve' to 'react-jsx' (React automatic runtime)
- Added '.next/dev/types/**/*.ts' to include paths
## MCP Configuration
- Removed unused context7 MCP server configuration
- Updated next-devtools command to use direct npx instead of cmd wrapper
## Verification
- Build: ✅ Successful (compiled in 2.8s with Turbopack)
- Lint: ✅ No errors or warnings
- TypeScript: ✅ Type checking passed
- Browser: ✅ App loads and runs correctly
This upgrade brings all the latest Next.js 16 features and improvements
including Turbopack as the default bundler, improved build performance,
and enhanced developer experience.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
This commit enhances the Model Context Protocol (MCP) server configuration
for improved cross-platform compatibility and adds additional MCP servers.
Key changes:
- Added separate MCP server configurations for Unix and Windows platforms
- shadcn-unix and shadcn-windows variants for proper command execution
- playwright-unix and playwright-windows variants for browser automation
- Integrated Context7 MCP server with API key authentication
- Updated package version from 1.1.10 to 1.1.11
- Applied configuration updates to both root .mcp.json and template
Technical details:
- Windows configurations use 'cmd /c' wrapper to properly execute npx commands
- Unix configurations use direct 'npx' command execution
- Context7 server uses HTTPS URL endpoint with header-based authentication
- Changes apply to both the main project and create-agentic-app template
Breaking changes: None
Migration notes: Users will need to set their CONTEXT7_API_KEY if using Context7
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Major improvements to AI integration and developer experience:
## OpenRouter Migration
- Replaced @ai-sdk/openai with @openrouter/ai-sdk-provider
- Updated chat API route to use OpenRouter for access to 100+ AI models
- Changed environment variables from OPENAI_* to OPENROUTER_*
- Updated default model from gpt-5-mini to openai/gpt-5-mini
- Benefits: Users can now access multiple AI providers (OpenAI, Anthropic, etc.) through a single unified API
## Documentation Overhaul
- Expanded CLAUDE.md from 3 lines to comprehensive 225-line guide
- Added complete project structure, tech stack overview, and AI integration guidelines
- Documented all environment variables, scripts, and common tasks
- Added critical rules for AI assistants working with the codebase
- Enhanced checkpoint command with detailed step-by-step instructions
## README Updates
- Updated feature list to highlight OpenRouter integration
- Revised setup instructions for OpenRouter API key acquisition
- Updated environment variable examples and deployment configuration
- Changed all OpenAI references to OpenRouter throughout documentation
## Create Command Updates
- Bumped create-agentic-app version from 1.0.0 to 1.0.1
- Synchronized all template files with main project documentation
- Updated template env.example and package.json with OpenRouter configuration
## Dependency Updates
- Added @openrouter/ai-sdk-provider@^1.2.0
- Updated @ai-sdk/openai from ^2.0.53 to ^2.0.60
- Updated @ai-sdk/react from ^2.0.78 to ^2.0.86
- Updated ai from ^5.0.78 to ^5.0.86
- Updated better-auth from ^1.3.29 to ^1.3.34
- Updated drizzle-kit from ^0.31.5 to ^0.31.6
- Updated eslint from ^9.38.0 to ^9.39.0
- Various type definition updates
Breaking Changes:
- Projects must update environment variables from OPENAI_* to OPENROUTER_*
- API keys must be obtained from openrouter.ai instead of platform.openai.com
- Model names now follow OpenRouter format (e.g., "openai/gpt-5-mini")
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
