automatasaurus 0.1.0

package/template/skills/notifications/SKILL.md

@@ -0,0 +1,95 @@
+---
+name: notifications
+description: System for alerting the user when attention is needed. Use when you need to get the user's attention for questions, approvals, or when stuck.
+---
+
+# Notification System
+
+Automatasaurus includes a notification system to alert the user when their attention is needed.
+
+## Automatic Notifications
+
+The system automatically sends notifications on stop based on context:
+- **Questions**: When Claude asks a question and waits for input
+- **Approvals**: When approval is needed to proceed
+- **Stuck**: When an agent encounters an issue
+- **Complete**: When all work is done
+
+## Explicit Notification Request
+
+When you need to explicitly alert the user, use:
+
+```bash
+.claude/hooks/request-attention.sh <type> "<message>"
+```
+
+### Notification Types
+
+| Type | When to Use | Sound |
+|------|-------------|-------|
+| `question` | You have a question that needs answering | Submarine |
+| `approval` | You need approval before proceeding | Submarine |
+| `stuck` | You've encountered an issue you can't resolve | Basso (alert) |
+| `complete` | All assigned work is finished | Hero (success) |
+| `info` | General notification | Glass |
+| `error` | An error occurred | Basso (alert) |
+
+### Examples
+
+```bash
+# Question needs answering
+.claude/hooks/request-attention.sh question "Should I use PostgreSQL or MySQL for this project?"
+
+# Approval needed
+.claude/hooks/request-attention.sh approval "PR #42 is ready for review"
+
+# Got stuck
+.claude/hooks/request-attention.sh stuck "Cannot access the GitHub API - authentication failed"
+
+# Work complete
+.claude/hooks/request-attention.sh complete "All 5 user stories have been implemented and tested"
+
+# Error occurred
+.claude/hooks/request-attention.sh error "Build failed with 3 errors"
+```
+
+## When to Notify
+
+### Always Notify For:
+- Questions that block progress
+- Security-related approvals
+- When you've been stuck for more than one attempt
+- Completion of significant milestones
+- Errors that require human intervention
+
+### Don't Notify For:
+- Minor progress updates
+- Self-recoverable errors
+- Questions you can answer from context
+
+## Configuration
+
+Notifications can be configured via environment variables:
+
+```bash
+# Disable sound
+AUTOMATASAURUS_SOUND=false
+
+# Custom log location
+AUTOMATASAURUS_LOG=/path/to/custom.log
+```
+
+## Platform Support
+
+- **macOS**: Native notifications with sound
+- **Linux**: Uses `notify-send` if available
+- **Windows**: PowerShell message box
+
+## Logging
+
+All notifications are logged to `$AUTOMATASAURUS_LOG` (default: `/tmp/automatasaurus.log`):
+
+```
+[2025-01-02 14:30:45] [question] Automatasaurus: Which database should we use?
+[2025-01-02 14:35:12] [complete] Automatasaurus: Feature implementation complete
+```

package/template/skills/pr-writing/SKILL.md

@@ -0,0 +1,259 @@
+---
+name: pr-writing
+description: Best practices for writing clear, effective pull request descriptions. Use when creating PRs to ensure they communicate changes effectively to reviewers.
+---
+
+# PR Writing Skill
+
+Guidelines for writing pull request descriptions that help reviewers understand and approve changes quickly.
+
+## PR Title Format
+
+```
+#<issue-number> <type>: <concise description>
+```
+
+Always prefix the PR title with the GitHub issue number to enable easy cross-referencing.
+
+### Types
+| Type | Use For |
+|------|---------|
+| `feat` | New feature or functionality |
+| `fix` | Bug fix |
+| `refactor` | Code restructuring without behavior change |
+| `docs` | Documentation only |
+| `test` | Adding or updating tests |
+| `chore` | Build, tooling, dependencies |
+| `perf` | Performance improvement |
+| `style` | Formatting, whitespace (no logic change) |
+
+### Good Titles
+```
+#42 feat: Add user registration endpoint
+#187 fix: Prevent duplicate form submissions
+#23 refactor: Extract authentication logic into service
+#91 docs: Add API documentation for user endpoints
+```
+
+### Bad Titles
+```
+Update code                    # Too vague, no issue number
+Fixed the bug                  # Which bug? No issue number
+WIP                           # Not ready for review
+feat: Add login               # Missing issue number prefix
+```
+
+## PR Body Structure
+
+```markdown
+## Summary
+[1-3 sentences explaining WHAT changed and WHY]
+
+Closes #<issue-number>
+
+## Changes
+- [Specific change 1]
+- [Specific change 2]
+- [Specific change 3]
+
+## Testing
+- [ ] Unit tests added/updated
+- [ ] Integration tests pass
+- [ ] Manual testing completed
+
+## Screenshots (if UI changes)
+[Before/after screenshots or GIFs]
+
+## Notes for Reviewers
+[Optional: Areas to focus on, known limitations, questions]
+```
+
+## Writing the Summary
+
+### Focus on WHY, not just WHAT
+
+**Bad:**
+> Added a new function called validateEmail and updated the user model.
+
+**Good:**
+> Adds email validation to prevent invalid emails from being stored, which was causing delivery failures in the notification system.
+
+### Be Specific About Impact
+
+**Bad:**
+> Fixed authentication bug.
+
+**Good:**
+> Fixes session timeout not being respected when "Remember Me" is checked, which was causing users to be logged out after 30 minutes regardless of their preference.
+
+### Keep It Scannable
+
+Reviewers often skim. Front-load the important information:
+
+**Bad:**
+> After investigating the issue that was reported last week about the dashboard loading slowly, I discovered that we were making redundant API calls. This PR addresses that by implementing caching.
+
+**Good:**
+> Reduces dashboard load time by 60% by caching API responses. The redundant calls were causing 3-second delays on each page load.
+
+## Describing Changes
+
+### Be Specific and Organized
+
+Group related changes:
+
+```markdown
+## Changes
+
+### API
+- Add `POST /api/users/register` endpoint
+- Add request validation middleware
+- Add rate limiting (10 requests/minute)
+
+### Database
+- Add `email_verified` column to users table
+- Add index on `email` column
+
+### Tests
+- Add unit tests for registration validation
+- Add integration tests for registration flow
+```
+
+### Explain Non-Obvious Changes
+
+```markdown
+## Changes
+- Increase connection pool size from 10 to 25
+  - Load testing showed connection exhaustion under peak traffic
+- Add retry logic to external API calls
+  - The payment provider has occasional 503s that resolve on retry
+```
+
+## Linking to Issues
+
+Always link the related issue:
+
+```markdown
+Closes #123
+```
+
+Or for partial work:
+
+```markdown
+Related to #123
+Part of #123
+```
+
+## Highlighting Review Focus
+
+Help reviewers prioritize their attention:
+
+```markdown
+## Notes for Reviewers
+
+**Please focus on:**
+- The caching invalidation logic in `src/cache/manager.ts` - I'm unsure if this handles all edge cases
+- The error handling in the new endpoint
+
+**You can skim:**
+- The test files follow existing patterns
+- The migration is straightforward
+```
+
+## Screenshots and Visuals
+
+For UI changes, always include:
+
+```markdown
+## Screenshots
+
+### Before
+[screenshot]
+
+### After
+[screenshot]
+```
+
+For complex flows, use GIFs or numbered screenshots:
+
+```markdown
+## User Flow
+
+1. User clicks "Register"
+   [screenshot 1]
+
+2. Validation errors shown inline
+   [screenshot 2]
+
+3. Success state with redirect
+   [screenshot 3]
+```
+
+## Breaking Changes
+
+Clearly call out breaking changes:
+
+```markdown
+## ⚠️ Breaking Changes
+
+- `getUserById()` now returns `null` instead of throwing when user not found
+- The `/api/v1/users` endpoint is deprecated, use `/api/v2/users`
+
+### Migration Guide
+1. Update all `try/catch` blocks around `getUserById()` to null checks
+2. Update API clients to use v2 endpoint
+```
+
+## Draft PRs
+
+Use draft PRs for:
+- Work in progress needing early feedback
+- Architectural decisions to discuss
+- Proof of concepts
+
+```markdown
+## 🚧 Draft PR
+
+This is a draft to get early feedback on the approach.
+
+**Questions:**
+1. Should we use Redis or in-memory caching?
+2. Is this the right place for this logic?
+
+**Not yet complete:**
+- [ ] Error handling
+- [ ] Tests
+- [ ] Documentation
+```
+
+## Checklist Before Submitting
+
+Before marking ready for review:
+
+- [ ] Title follows `#<issue> type: description` format
+- [ ] Summary explains what AND why
+- [ ] Issue is linked with `Closes #X`
+- [ ] Changes are listed specifically
+- [ ] Tests are included and passing
+- [ ] Screenshots added for UI changes
+- [ ] Breaking changes are highlighted
+- [ ] Self-reviewed the diff for obvious issues
+- [ ] Branch is synced with main (no merge conflicts)
+
+## PR Size Guidelines
+
+| Size | Lines Changed | Review Time | Recommendation |
+|------|---------------|-------------|----------------|
+| Small | < 100 | < 30 min | Ideal |
+| Medium | 100-300 | 30-60 min | Acceptable |
+| Large | 300-500 | 1-2 hours | Consider splitting |
+| XL | > 500 | Hours | Split if possible |
+
+If a PR is large, explain why:
+```markdown
+## Note on PR Size
+
+This PR is larger than usual because it includes a database migration
+that requires updating all affected queries atomically. Splitting would
+leave the codebase in a broken state.
+```

package/template/skills/project-commands/SKILL.md

@@ -0,0 +1,100 @@
+---
+name: project-commands
+description: Manages project-specific commands for development, testing, and deployment. Use when needing to run project commands or when setting up commands for a new project.
+---
+
+# Project Commands Skill
+
+This skill helps agents find and use the correct commands for the current project.
+
+## Finding Commands
+
+All project-specific commands are defined in `.claude/commands.md`. Always read this file before running commands.
+
+```bash
+# Read the commands file
+cat .claude/commands.md
+```
+
+## Command Categories
+
+### Development
+- **Install**: Set up project dependencies
+- **Dev**: Start development server
+- **Build**: Create production build
+
+### Testing
+- **Test**: Run all tests
+- **Test:unit**: Unit tests only
+- **Test:e2e**: End-to-end tests (Playwright)
+- **Test:coverage**: Tests with coverage report
+
+### Code Quality
+- **Lint**: Check code style
+- **Format**: Auto-format code
+- **Typecheck**: Verify types
+
+## Setting Up Commands for a New Project
+
+1. Copy the template:
+```bash
+cp .claude/commands.template.md [target-project]/.claude/commands.md
+```
+
+2. Replace placeholders with actual commands:
+```
+{{install}} → npm install (or yarn, pnpm, etc.)
+{{dev}} → npm run dev
+{{dev_url}} → http://localhost:3000
+{{test}} → npm test
+```
+
+3. Remove unused sections (database, docker, etc.)
+
+## Common Patterns by Stack
+
+### Node.js / npm
+```
+{{install}}: npm install
+{{dev}}: npm run dev
+{{test}}: npm test
+{{build}}: npm run build
+```
+
+### Python / pip
+```
+{{install}}: pip install -r requirements.txt
+{{dev}}: python manage.py runserver
+{{test}}: pytest
+{{build}}: python setup.py build
+```
+
+### Go
+```
+{{install}}: go mod download
+{{dev}}: go run .
+{{test}}: go test ./...
+{{build}}: go build
+```
+
+### Rust
+```
+{{install}}: cargo build
+{{dev}}: cargo run
+{{test}}: cargo test
+{{build}}: cargo build --release
+```
+
+## Agent Usage
+
+When an agent needs to run a command:
+
+1. Check `.claude/commands.md` for the correct command
+2. Use the command exactly as specified
+3. If a command is missing, ask the user or check `package.json`/equivalent
+
+## Updating Commands
+
+If you discover the correct command for an action:
+1. Update `.claude/commands.md` with the correct command
+2. Ensure other agents can find it in the future