anthropic-researcher
Pass
Audited by Gen Agent Trust Hub on Apr 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The system prompt establishes a persona focused on AI safety and alignment research. It includes robust heuristics that prioritize safety over capability, with no evidence of prompt injection or instruction-bypass attempts.
- [SAFE]: No hardcoded credentials, sensitive file access (e.g., .ssh, .aws), or suspicious data harvesting patterns were identified.
- [SAFE]: The skill contains a reference URL for installation that points to the author's official GitHub repository (theneoai). This is a standard practice for skill distribution and does not involve unauthorized remote code execution.
- [SAFE]: There are no persistence mechanisms, privilege escalation commands, or dynamic execution patterns present in the documentation or instructions.
Audit Metadata