Documentation techniquesource GitHub
Créer des TODO intelligents
/create-todosJ’analyse les opérations récentes et j’ajoute des commentaires TODO contextuels dans votre code.
// contenu du skill
Create Smart TODOs
I'll analyze recent operations and create contextual TODO comments in your code.
First, let me understand your project standards:
- Read README.md for project conventions
- Read CONTRIBUTING.md for code style guidelines
- Grep existing TODOs to match your format
- Read documentation for technical context
I'll examine what just happened in our session:
- Security scan findings that need fixes
- Test failures requiring attention
- Code review issues to address
- Architecture improvements identified
- Performance optimizations suggested
Using native tools to analyze context:
- Grep tool to find related code sections
- Read tool to understand implementation details
- TodoWrite to track what TODOs I'm creating
For each TODO I'll:
- Find the exact location where it belongs
- Analyze surrounding code for context
- Create clear, actionable TODO comments
- Include priority and category markers
TODO Format Adaptation:
I'll match your project's existing TODO style:
- Check if you use categories like
[Security]or(SECURITY) - Match comment style:
//,#,/* */, etc. - Follow any ticket reference patterns your project uses
- Respect line length limits from your linter
Smart Context References:
I'll create TODOs that reference the source of the issue:
- Security findings will reference the scan line number
- Performance issues will include measured thresholds
- Bug fixes will reference failing test locations
- Refactoring needs will reference architectural patterns
I'll use intelligent placement:
- Security TODOs near vulnerable code
- Performance TODOs at bottleneck locations
- Bug fixes where errors occur
- Architecture TODOs at integration points
After creating TODOs, I'll ask: "How would you like to track these?"
- Convert to GitHub issues with
/todos-to-issues - Keep as code comments for gradual resolution
- Generate summary report for team review
Important: I will NEVER:
- Add "Generated by Claude" or AI attribution in TODOs
- Create vague or non-actionable TODOs
- Place TODOs in random locations
- Overwhelm code with excessive comments
This creates a clear technical debt roadmap directly in your codebase.
// source originale publique
notlikeDev/CCPlugins/commands/create-todos.md
Licence : MIT License
Projet indépendant, non affilié à Anthropic. Ce skill reste la propriété de son auteur original.