The Agent Skills Directory

No Executable Content: The skill is composed entirely of markdown-based instructions and does not ship with any scripts, binary files, or system commands.
Indirect Prompt Injection Surface: The skill is designed to process and analyze external data such as code snippets, PRs, and diffs (Ingestion point: SKILL.md). While there are no explicit boundary markers or sanitization instructions mentioned (Absent), the skill's capability inventory shows no access to the file system, network, or code execution environments (Capability inventory: None). This lack of privileges ensures that even if malicious instructions were present in the analyzed data, the agent would have no means to execute them, rendering the surface safe.

code-review