poll-github-work
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted external data from GitHub issues, pull requests, and comments, creating a surface for indirect prompt injection attacks.
- Ingestion points: The skill reads external content using
gh issue viewandgh pr viewcommands to fetch bodies and comments from GitHub (SKILL.md). - Boundary markers: The instructions include a 'Poll Safety' section explicitly warning the agent to treat this content as untrusted data and to ignore any instructions found within it.
- Capability inventory: The skill possesses the ability to read comprehensive repository metadata and, upon explicit user selection, mutate the repository state by creating branches and pull requests.
- Sanitization: The skill relies on behavioral instructions rather than technical sanitization or delimiters to handle untrusted input.
Audit Metadata