kill-dead-code
Pass
Audited by Gen Agent Trust Hub on Jul 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill's instructions are focused on standard code maintenance tasks such as searching for symbol references, deleting unused code, and running build scripts to verify integrity.
- [SAFE]: Indirect Prompt Injection Surface. The skill instructs the agent to read repository code via grep and execute build commands. While this is an attack surface, it is standard for code-editing tools and the instructions do not encourage unsafe behavior. Mandatory evidence: 1. Ingestion: Symbol searching via grep (SKILL.md). 2. Boundaries: Absent. 3. Capabilities: Code deletion and build execution (SKILL.md). 4. Sanitization: Absent.
Audit Metadata