Automated development workflow with task board management, code quality checks, and continuous improvement. Perfect for auto-coding projects every 30 minutes.
---
name: auto-improve
description: "Automated development workflow with task board management, code quality checks, and continuous improvement. Perfect for auto-coding projects every 30 minutes."
metadata:
{
"openclaw":
{
"requires": { "bins": ["git", "python3"] },
"install": [],
},
}
---
# Auto-Improve Skill
Automated development workflow with task board management, code quality checks, and continuous improvement.
## Features
- **Task Board Management**: Read and update TASKS.md for task tracking
- **Multi-language Support**: Python, JavaScript, TypeScript, and more
- **Code Quality**: Syntax checking, static analysis, security scanning
- **Deployment Testing**: Automated testing before commit
- **Continuous Improvement**: Learn from each iteration
- **Error Handling**: Automatic retry mechanism for API rate limits
## Usage
This skill provides a comprehensive automated development workflow that runs every 30 minutes (configurable). It's perfect for:
- Auto-coding projects
- Continuous development
- Task-driven development
- Quality assurance automation
## Workflow
1. **Preparation**: Git sync and branch management
2. **Task Selection**: Read TASKS.md and pick next task
3. **Implementation**: Code the solution
4. **Quality Check**: Syntax, linting, security
5. **Deployment Test**: Test before commit
6. **Commit & Push**: Version control
7. **Task Update**: Mark task as complete
8. **Report**: Generate completion report
## Error Handling
This skill includes robust error handling for API rate limits and other transient errors:
### Automatic Retry Mechanism
```markdown
## 错误处理机制
如果遇到 API 速率限制或其他错误:
1. 等待 5 分钟后重试
2. 如果重试失败,等待 15 分钟后再次重试
3. 如果仍然失败,记录错误并跳过本次执行
4. 不要因为错误而停止整个任务
```
### Best Practices
- **Extended Timeout**: Use 30-minute timeout for complex tasks
- **Error Logging**: Record errors for debugging
- **Graceful Degradation**: Skip failed runs, continue on next schedule
- **No Stop on Error**: Keep the task running even if individual runs fail
## Configuration
The workflow can be customized by modifying the task prompt in your cron job. Key configurable areas:
- **Project path**: Where your code lives
- **Task board**: TASKS.md location and format
- **Quality tools**: Which linters/analyzers to use
- **Testing strategy**: How to test before commit
- **Git settings**: Remote URLs, branch strategy
- **Error handling**: Retry logic and timeout settings
## Task Board Format
Create a `TASKS.md` file in your project root:
```markdown
# Project Tasks
## P0 - MVP (Must Have)
- [ ] User authentication system
- [ ] Database models
- [ ] API endpoints
## P1 - Important
- [ ] Input validation
- [ ] Error handling
- [ ] Logging system
## P2 - Nice to Have
- [ ] Performance optimization
- [ ] Documentation
- [ ] Unit tests
```
## Best Practices
1. **Always test before commit** - Deploy and verify functionality
2. **Small, focused tasks** - Each task should be completable in one cycle
3. **Clear task descriptions** - Include enough detail for implementation
4. **Regular reviews** - Periodically review and update task priorities
5. **Backup strategy** - Keep backups of important configurations
6. **Error handling** - Implement retry logic for transient failures
## Security Notes
- Store sensitive credentials securely (not in TASKS.md)
- Use environment variables for configuration
- Regular security scans with bandit/safety
- Keep dependencies updated
## Troubleshooting
### Common Issues
1. **Git authentication failed**: Check remote URL and credentials
2. **Task not found**: Verify TASKS.md format and location
3. **Quality checks fail**: Review linter output and fix issues
4. **Deployment test fails**: Check service dependencies and ports
5. **API rate limit**: Implement retry logic with exponential backoff
### Debug Mode
Add verbose logging to your workflow for debugging:
```bash
set -x # Enable debug output
```
## Contributing
This skill is designed to be extended. Common customizations:
- Add new quality tools
- Support additional languages
- Customize testing strategies
- Add notification systems
- Enhance error handling
## License
MIT License - Feel free to modify and distribute.don't have the plugin yet? install it then click "run inline in claude" again.