The Agent Skills Directory

[SAFE]: The skill provides comprehensive engineering guidelines that prioritize software reliability and agent efficiency. It uses industry-standard tool recommendations (e.g., Rust, TypeScript strict mode, Ruff) and cites legitimate research and security reports.
[SAFE]: It enforces a clear hierarchy of authority where explicit user instructions and project-specific configurations override the skill's defaults, preventing unintended behavioral lock-in.
[SAFE]: Security best practices are explicitly integrated, such as mandatory TDD, input validation at boundaries using Zod/Pydantic, and the inclusion of security scanners like gitleaks and bun audit in the development lifecycle.
[SAFE]: The skill author references their own project resources and well-known technology providers in a transparent manner for benchmarking and tool evaluation purposes.
[SAFE]: Instructions for agents to inspect the live state of the environment (e.g., using .schema or ls) are geared toward reducing hallucinations and improving task accuracy, rather than being used for malicious reconnaissance.

coding-standards