implement
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes content fetched from external issue trackers to guide its implementation tasks. This data is untrusted and could contain embedded instructions aimed at manipulating the agent's output or actions.
- Ingestion points: External issue tracker content referenced in SKILL.md.
- Boundary markers: None specified; the agent is not instructed to treat ticket content as data separate from its primary instructions.
- Capability inventory: The skill possesses significant capabilities, including file modification, execution of local tests/typecheckers, and committing code to branches via the /skill:commit-buddy reference.
- Sanitization: The instructions lack requirements for sanitizing or validating external ticket content before processing it.
Audit Metadata