audit-code-reviewer
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) as it processes potentially untrusted content from git diffs and plan documents and passes them to subagents for analysis. Ingestion points: The skill gathers context from
git diff,git status, user requests, and plan documents (SKILL.md). Boundary markers: The prompt template used to dispatch subagents lacks explicit instructions or markers to isolate the code data and prevent the model from following any embedded instructions found within the code changes (SKILL.md). Capability inventory: The skill has the authority to directly apply 'low-risk fixes' to the file system, which provides a pathway for an injection attack to result in unauthorized file modifications (SKILL.md). Sanitization: There is no documented process for sanitizing or escaping the git diff content before it is processed. - [COMMAND_EXECUTION]: The workflow involves executing local shell commands including
git status,git diff, and various 'focused test commands' to evaluate the repository state and the impact of changes (SKILL.md).
Audit Metadata