The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the execution of local development commands such as npm run build, pytest, and ruff check. These are standard tools used for code verification within a developer environment.
[DATA_EXPOSURE]: Performs local security scans using grep to find hardcoded credentials like API keys (sk-, api_key) and debug statements (console.log, print) to prevent accidental exposure in production or pull requests. This is a defensive security feature.
[PROMPT_INJECTION]: The skill processes untrusted data from the local filesystem (source code, git diffs) and tool outputs to generate reports. This creates a potential surface for indirect prompt injection if malicious instructions are embedded in the code comments or logs. Ingestion points: project files, git diff output; Boundary markers: absent; Capability inventory: shell access for build/test tools; Sanitization: absent. Severity is assessed as low.
[EXTERNAL_DOWNLOADS]: References well-known security auditing tools like pip-audit. These are considered safe as they target official package registries.

verification-loop