refactor-security
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates in a read-only, advisory capacity and does not attempt to execute the code it audits or perform any network operations. It is a defensive tool focused on security analysis.\n- [PROMPT_INJECTION]: The skill is designed to ingest and process untrusted code diffs, which represents a potential surface for indirect prompt injection. However, the skill provides a structured methodology and explicit guardrails (e.g., 'advisory-only', 'edits nothing') to mitigate the risk of an agent being misled by instructions embedded in the audited code. No evidence of malicious injection was found within the skill files themselves.\n
- Ingestion points: The skill reads external code changes via git diff and file system access (SKILL.md, references/method.md).\n
- Boundary markers: The instructions do not define specific delimiters to isolate the untrusted code content from the agent's operational instructions.\n
- Capability inventory: The skill utilizes shell commands to retrieve diffs and executes a local Node.js script (scripts/checklist.mjs) to provide an audit checklist.\n
- Sanitization: No explicit sanitization or filtering of the audited code content is performed before processing.
Audit Metadata