The Agent Skills Directory

[SAFE]: The skill's primary purpose is to assist in code quality and security verification. Its components, including the automated verification script and reference guides, align with this stated goal.
[COMMAND_EXECUTION]: The skill utilizes a bash script (Scripts/verify-pipeline.sh) to execute standard development tools like npm, cargo, and pytest. These commands are hardcoded within the script and do not accept untrusted external input, minimizing the risk of command injection.
[INDIRECT_PROMPT_INJECTION]: As a code review tool, the skill naturally ingests untrusted data in the form of the codebase it is assigned to analyze. While there is a theoretical risk that malicious comments or code within a project could attempt to influence the agent's behavior, the skill provides clear internal guidelines for the agent to maintain objective verification standards.
Ingestion points: Reads source code files via Read, Grep, and Glob tools.
Boundary markers: Not explicitly defined in the provided instruction set.
Capability inventory: Includes Bash for running tests, Write for applying refactors, and Browser for UI auditing.
Sanitization: Relies on the agent's instructional logic and the provided Security Verification Guide to identify and mitigate risks during the review process.

code-verify-and-clean-skill