human-readable-code-ja
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references established programming guidelines and documentation from well-known technical domains including GitHub, Go.dev, and Python.org for informational purposes.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied code and project metadata to perform its code improvement tasks, presenting a potential surface for indirect injection.
- Ingestion points: The skill reads and interprets code, project instructions, and existing comments provided in the agent context as described in
SKILL.md. - Boundary markers: The instructions recommend prioritizing existing project policies and explicit designations, though they do not define specific technical delimiters for untrusted input.
- Capability inventory: The skill provides logic for code generation and modification based on user input, which may be part of multi-step tool chains involving file read/write operations.
- Sanitization: The guidelines focus on stylistic and structural improvements and do not explicitly detail sanitization or escaping steps for the input data.
Audit Metadata