review-changes
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted content from the repository's code diffs and commit messages, which could contain instructions designed to influence the agent's behavior.
- Ingestion points: Output from
git diff,git log, andgit showas specified inSKILL.mdStep 2. - Boundary markers: The instructions lack explicit delimiters or warnings to isolate the diff content from the agent's operational logic.
- Capability inventory: The agent executes shell commands (
git) and provides evaluative feedback to the user. - Sanitization: No sanitization or filtering logic is defined for the repository content being reviewed.
- [COMMAND_EXECUTION]: The skill is designed to execute shell commands to inspect the repository state.
- Evidence: Step 2 in
SKILL.mdlists several commands includinggit status,git branch,git log,git diff, andgit show. - Description: The skill constructs these commands using scopes (branches, commits, or file paths) provided by the user. This creates an attack surface for command injection if the agent fails to properly normalize or validate these strings before execution.
Audit Metadata