The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its core functionality of processing external review data.
Ingestion points: Pull request comments and code reviews are retrieved via gh pr view and source files are analyzed using grep as specified in the SKILL.md instructions.
Boundary markers: No explicit delimiters or instructions are provided to distinguish between the agent's system prompts and instructions that might be embedded within the review data.
Capability inventory: The agent has access to Bash, Write, and Edit tools, which could be exploited for unauthorized actions if the agent follows malicious commands found in the input data.
Sanitization: The skill lacks any evidence of sanitization or validation for the content retrieved from external sources before it is processed.

agent-peer-review-checker