medical-imaging-review
Pass
Audited by Gen Agent Trust Hub on May 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [SAFE]: Analysis of the instructions and reference files shows no evidence of malicious patterns, bypass attempts, or obfuscation.
- [EXTERNAL_DOWNLOADS]: The skill retrieves metadata and full-text papers from reputable sources such as ArXiv, PubMed, and Crossref.
- [COMMAND_EXECUTION]: Includes documentation for user-led configuration of Model Context Protocol (MCP) servers and Zotero API access.
- [PROMPT_INJECTION]: Ingestion of external research papers represents a potential surface for indirect prompt injection. 1. Ingestion points: Abstracts and full-text from ArXiv, PubMed, and Zotero. 2. Boundary markers: Absent. 3. Capability inventory: Local file writing for manuscript generation and network requests for academic metadata. 4. Sanitization: Absent.
Audit Metadata