code-simplifier
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions and associated reference files (refactoring patterns) focus exclusively on improving code quality. All logic is static and instructional, with no implementation of automated execution or dangerous shell commands.
- [SAFE]: Metadata and licensing information are transparent. The author links to their GitHub profile, and the license is a standard Creative Commons variant with specific permissions, posing no security threat.
- [INDIRECT_PROMPT_INJECTION]: The skill presents a data ingestion surface as it is designed to analyze and refactor external code and pull request content.
- Ingestion points: Code snippets or PR content provided via user input or the agent's file system tools (e.g.,
read_file). - Boundary markers: None explicitly defined in the prompt instructions to isolate untrusted data.
- Capability inventory: The skill itself does not invoke tools, but it guides an agent that typically has file system read/write access and shell execution capabilities.
- Sanitization: None present. However, the risk is negligible as the skill instructs the agent to simplify code structure rather than execute or interpret its logic as instructions, and no dangerous interpolation patterns were observed.
Audit Metadata