refactor
Pass
Audited by Gen Agent Trust Hub on Jul 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is focused on code refactoring and uses local context such as git history and file contents to perform its tasks.
- [COMMAND_EXECUTION]: The instructions direct the agent to run standard development commands including
git diffand local test suites or verification scripts. These operations are restricted to the local environment and are necessary for the skill's primary purpose of code verification. - [PROMPT_INJECTION]: While the skill ingests external data (source code and git diffs) which could contain indirect prompt injection attempts, it includes specific rules for behavior preservation and verification that act as guardrails against malicious instructions embedded in the code being refactored.
Audit Metadata