The Agent Skills Directory

[SAFE]: The skill provides structured instructions for a development workflow without containing prompt injections, behavioral overrides, or safety bypasses.
[COMMAND_EXECUTION]: The skill facilitates running local tests by identifying project-specific runners (e.g., in package.json, Cargo.toml, or pyproject.toml). Execution is limited to the local environment using the project's own tooling.
[SAFE]: The skill reads local project files and configuration to understand the development environment and test setup, which is necessary for its function. No evidence of unauthorized data exfiltration was detected.
[PROMPT_INJECTION]: The skill uses $ARGUMENTS to receive user feature descriptions but does not include any instructions to override core agent rules or bypass constraints.

tdd