multi-omics-clinical-integration-planner
Pass
Audited by Gen Agent Trust Hub on Aug 6, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill instructions maintain clear boundaries for the AI's behavior, specifically prohibiting the generation of fabricated medical datasets, PMIDs, or DOIs. No bypass or override patterns were detected.- [DATA_EXFILTRATION]: No sensitive file paths or unauthorized network operations were found. The skill only includes a reference to the author's GitHub repository for documentation purposes.- [REMOTE_CODE_EXECUTION]: There is no evidence of remote script downloading, package installation, or dynamic code execution. The skill is entirely text-based.- [SAFE]: Independent analysis of all markdown instructions, reference modules, and evaluation metadata confirms the absence of obfuscation, persistence mechanisms, or privilege escalation. The self-referential safety claims within the included evaluation report were identified as non-authoritative data points and did not influence the final verdict.
Audit Metadata