clean-code-patterns
Pass
Audited by Gen Agent Trust Hub on May 5, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to analyze and refactor code, which creates a surface for indirect prompt injection if the processed content contains hidden instructions.\n
- Ingestion points: External code accessed via the Read tool during review tasks as described in SKILL.md.\n
- Boundary markers: None identified; there are no instructions to delimit or ignore instructions embedded within the code being analyzed.\n
- Capability inventory: The skill lists Read, Write, Edit, and Grep tools in its frontmatter, enabling the agent to modify files based on its analysis.\n
- Sanitization: No sanitization or content validation steps are specified for the input code.\n- [NO_CODE]: The skill consists entirely of instructional text and does not include any executable scripts, binaries, or automated tasks.
Audit Metadata