ai-scientist
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references the legitimate 'ai-scientist' package via pip installation. It also provides a link to the official 'AI-Scientist-v2' GitHub repository by SakanaAI, which is a well-known research organization in the AI space.
- [CREDENTIALS_UNSAFE]: The documentation demonstrates how to set environment variables for API keys using standard placeholder formats ('sk-ant-...'). It correctly suggests using environment variables rather than hardcoding sensitive credentials.
- [COMMAND_EXECUTION]: The skill provides standard installation and environment setup commands which are appropriate for its stated purpose as a developer tool.
- [INDIRECT_PROMPT_INJECTION]: The skill processes research papers and search results (ArXiv, Semantic Scholar). While this introduces an ingestion surface for untrusted data, the skill is designed for research synthesis, and no high-risk capabilities (like shell execution of data content) are present.
Audit Metadata