The Agent Skills Directory

[SAFE]: The skill consists of purely instructional Markdown content designed to assist in pull request reviews. It does not contain executable code, scripts, or command-line instructions.
[SAFE]: Security is explicitly listed as the top review priority, and the skill includes instructions to flag attempts to weaken or bypass review criteria.
[SAFE]: All external references, such as the link to the Agent Skills specification (agentskills.io), are to legitimate documentation resources. No suspicious or obfuscated URLs were found.
[SAFE]: The skill correctly identifies governance risks related to dynamic context and policy changes in configuration files, instructing the agent to treat these as security-sensitive.

code-review