The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute standard development commands (e.g., npm run build, npx tsc, ruff check, npm run test) to verify code quality. These are expected for a verification skill.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill includes grep commands to look for hardcoded secrets (sk-, api_key) in the local codebase. This is a security best practice for code review and does not involve network exfiltration.
[SAFE]: The operations described are standard software engineering practices. The skill uses well-known local tools (TypeScript compiler, Ruff, Grep, Git) and does not reference suspicious external domains or execute remote scripts.

verification-loop