rc-triage
Pass
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to reproduce bugs by tracing code and running 'tests or commands' based on a reporter's steps. This creates a risk where the agent might execute malicious commands embedded in an issue description or comment by an untrusted user.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing external, untrusted content from issue trackers and using it to drive its state machine and reproduction logic.
- Ingestion points: Reads issue bodies, comments, and reporter details from
SKILL.md(via the issue tracker interface). - Boundary markers: No specific delimiters or safety instructions are provided to help the agent distinguish between the triage instructions and potentially malicious instructions hidden within the issue content.
- Capability inventory: The agent is authorized to run shell commands/tests and write markdown files to the
.out-of-scope/directory. - Sanitization: There is no explicit requirement for the agent to sanitize or validate content extracted from the issue tracker before acting on it.
Audit Metadata