deslop
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions are focused on code quality and style maintenance. There are no signs of malicious commands, sensitive data access, or network exfiltration.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data (code diffs), which inherently creates a surface for indirect prompt injection. However, as a development tool intended for code review, this is a standard operational pattern. The instructions emphasize structural cleanup and simplification, which does not involve executing logic found in the analyzed code.
- Ingestion points: Code diffs (SKILL.md)
- Boundary markers: None specified
- Capability inventory: Reading and writing code files
- Sanitization: None specified
Audit Metadata