afk
Pass
Audited by Gen Agent Trust Hub on Jul 25, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes instructions from an external issue tracker. 1. Ingestion points: Issue descriptions and tracker metadata used to identify tasks. 2. Boundary markers: No delimiters or specific instructions to ignore embedded directives within the issue body are provided. 3. Capability inventory: The agent can create branches (/wt-switch-create), commit code (/commit), use a browser (/agent-browser), and open pull requests (/open-pr). 4. Sanitization: No sanitization of the issue content is specified, though the skill includes a code review step and human-in-the-loop stop conditions.
Audit Metadata