implement
Pass
Audited by Gen Agent Trust Hub on Jul 14, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted external data (PRDs and issues) to drive implementation and commits, creating a surface where malicious instructions embedded in these documents could potentially influence agent behavior.\n
- Ingestion points: User-provided PRDs or issue descriptions (SKILL.md).\n
- Boundary markers: Absent. The instructions do not define clear delimiters or warnings for the agent to ignore directives within the input data.\n
- Capability inventory: The skill describes a workflow involving code implementation, execution of typechecking and testing tools, and git commits.\n
- Sanitization: Absent. No explicit validation or filtering of the input documentation content is required before the agent acts upon it.
Audit Metadata