nw-refactor
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface because it ingests untrusted source code from the project directory. If the code contains malicious comments or instructions, it could attempt to manipulate the agent's refactoring logic.
- Ingestion points: The skill reads from src/* and tests/* for context.
- Boundary markers: No explicit delimiters are used to separate untrusted code from instructions.
- Capability inventory: The skill can modify production code, test files, and execute the project's test suite via command execution.
- Sanitization: No sanitization of code content is performed. This risk is inherent to code-editing tasks and is mitigated by the skill's requirement to run and pass the existing test suite.
- [COMMAND_EXECUTION]: The skill is instructed to run the local test suite exactly once after applying a batch of refactors. This is an expected and necessary operation for the skill's purpose of code improvement and verification.
Audit Metadata