ai-researcher
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a purely instructional expert profile (persona) designed to guide the AI's reasoning and methodology in machine learning research.
- [SAFE]: Analysis of the markdown content for prompt injection, data exfiltration, and obfuscation yielded no findings. The mention of 'prompt injection' in the text refers to a recommended security evaluation protocol for researchers to follow, rather than an attempt to override agent behavior.
- [SAFE]: No automated command execution or dynamic context injection patterns were detected. All mentioned tools (e.g., PyTorch, JAX, Hugging Face) are part of a descriptive catalog of standard industry resources.
- [SAFE]: No hardcoded credentials or sensitive file path accesses were identified.
Audit Metadata