This is a personal repository used to codify my engineering workflows into reusable AI agent skills.
Most AI agents struggle with context limits. If you feed an LLM a massive task with too many details, it hits a context cut-off and loses track of the goal. To solve this, this repository treats AI skills like tasks in an Agile sprint: as small, specific, and particular as possible.
| Skill | Location | Description |
|---|---|---|
| Prop Drilling Detector | skills/react/detect-prop-drilling/SKILL.md |
Analyzes React component trees to identify props passed through multiple layers without being used, and suggests Context API or composition-based refactors. |
| React Declarative Analyzer | skills/react/react-declarative-analyzer/SKILL.md |
Scans React components for imperative patterns such as manual DOM work or unnecessary effects and recommends declarative state-driven alternatives. |
| React SRP/DRY Analyzer | skills/react/react-srp-dry-analyzer/SKILL.md |
Reviews components for SRP and DRY violations and suggests smaller components or hooks to improve composition and reuse. |
| Test Failure Resolver | skills/ci/test-failure-resolver/SKILL.md |
Helps trace failing tests by inspecting dependencies and recent changes, then routes the issue to the most relevant developer without blame. |
- Add Codex/OpenAI agent metadata to each skill so they are discoverable and consistent across tooling.
To ensure users can easily pull these skills down via CLI, use the following command:
npx skills@latest add jericbas/skills
npx skills@latest add jericbas/skills -a github-copilot
npx skills@latest add jericbas/skills -a claude-code