refactoring
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is entirely composed of instructional markdown content designed to guide an agent through a safe refactoring process. No executable scripts, subprocess calls, or network operations are present.
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process and modify external codebases, which constitutes a data ingestion surface. 1. Ingestion points: Source code files, surrounding code context, and dependencies (SKILL.md). 2. Boundary markers: The skill establishes logical boundaries by requiring the preservation of existing behavior, return types, and side effects. 3. Capability inventory: The guided agent is expected to perform file writes and execute verification tests or linters. 4. Sanitization: There is no automated sanitization of the input code; the skill relies on the agent's compliance with safety instructions to mitigate risks from malicious code patterns.
Audit Metadata