refactoring
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists entirely of instructional content in markdown format providing best practices for code refactoring. It does not include any executable code, scripts, or requests for sensitive information.
- [INDIRECT_PROMPT_INJECTION]: The skill describes a process for the agent to ingest and analyze external data like source code and code-review reports. While this establishes an indirect prompt injection attack surface, the workflow mitigates risk by requiring the agent to maintain strict behavioral invariance and verify all changes through a comprehensive test suite.
- Ingestion points: Local source code files and external code-review reports.
- Boundary markers: None explicitly defined.
- Capability inventory: The skill assumes the ability to execute local build and test commands via agent tools.
- Sanitization: Relies on automated test verification and manual/automated diffing of public interfaces to prevent unintended behavioral changes.
Audit Metadata