implement-fanout
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection because it retrieves task descriptions from an external file to dynamically generate instructions for subordinate agents.
- Ingestion points: Reads ticket content and dependency data from 'docs/agents/issue-tracker.md'.
- Boundary markers: The skill logic does not specify the use of delimiters or 'ignore embedded instructions' warnings when passing ticket data to subagents.
- Capability inventory: The skill can spawn new subagents, perform git operations (commit, push), and create pull requests via the agent harness.
- Sanitization: There is no mention of sanitizing or validating the contents of the tickets before they are interpolated into the briefs for the subagents.
Audit Metadata