The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting untrusted data from GitHub PR comments and using it to guide code modifications.\n- Ingestion points: In SKILL.md, the skill uses gh api graphql to fetch content from reviewThreads and reviews on GitHub.\n- Boundary markers: The instructions lack delimiters or system-level warnings to distinguish between legitimate code review feedback and malicious instructions embedded in comments.\n- Capability inventory: The skill employs the Bash, Edit, and Write tools, and can perform git push operations, providing a significant impact path for injected instructions.\n- Sanitization: There is no evidence of content validation or sanitization of the fetched review comments before they are processed by the agent to implement fixes.

address-review