implement
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided specs and tickets in SKILL.md, which are untrusted data sources. It lacks boundary markers to delimit user input from instructions. The agent has capabilities to run git commands and external verification tools. No sanitization is mentioned for the input.
- [COMMAND_EXECUTION]: Instructions in SKILL.md direct the agent to execute git revision tracking, git commits, and verification tools like test-audit, TDD tools, and code-review tools.
Audit Metadata