code-simplification
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions define a clear, restricted scope for code refactoring focused on clarity and intent without introducing new abstractions or architectural rewrites.
- [PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it ingests and processes external source code and diffs.
- Ingestion points: The skill reads changed code and diffs in context from the project directory as specified in SKILL.md.
- Boundary markers: The contract explicitly instructs the agent to preserve security and data-integrity boundaries, which serves as a safety constraint for handling external content.
- Capability inventory: The skill is authorized to read files, modify source code, and execute project-specific checks or tests.
- Sanitization: No explicit code-level sanitization is defined, though the process is gated by a requirement that the initial candidate state must be GREEN.
Audit Metadata