The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses standard instructional language to guide the agent through a code review workflow. No attempt to override safety guidelines, bypass constraints, or extract system prompts was detected.
[DATA_EXPOSURE]: No hardcoded credentials, secrets, or sensitive file paths were found. The skill does not perform any network operations to external or untrusted domains.
[REMOTE_CODE_EXECUTION]: The skill is entirely composed of Markdown documentation and instructions. It does not contain any executable scripts, package installation commands, or remote code download patterns.
[OBFUSCATION]: No obfuscated content, such as Base64-encoded strings, zero-width characters, or homoglyph attacks, was detected in the skill files.
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to process external code changes (pull requests), it includes explicit safety guidelines, such as 'Treat model metadata, file contents, and user-provided values as untrusted' and 'Check bounds, overflow, signed/unsigned conversions'. This demonstrates a secure-by-design approach to handling untrusted input.
[DYNAMIC_CONTEXT_INJECTION]: No use of dynamic context execution patterns (e.g., shell commands executed at load time) was found.

code-review