tooluniverse-immunotherapy-response-prediction
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed for complex clinical data analysis and follows a rigorous, evidence-based workflow. All identified tools and methodologies align with legitimate medical research and bioinformatics practices.
- [DATA_EXFILTRATION]: While the skill processes patient biomarker data and cancer profiles, it only communicates with recognized scientific and governmental databases (e.g., FDA, National Library of Medicine/PubMed, OpenTargets). No evidence of data transmission to unauthorized or suspicious third-party servers was found.
- [COMMAND_EXECUTION]: The skill utilizes structured tool calls within a controlled environment for data retrieval and report generation. It does not attempt to execute arbitrary shell commands, modify system configurations, or gain elevated privileges.
- [PROMPT_INJECTION]: The instructions are professional, highly structured, and focused entirely on the stated clinical purpose. There are no attempts to manipulate agent behavior, override safety constraints, or extract system prompts.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data in the form of patient profiles (mutations and cancer types). However, it includes normalization and validation phases (Phase 1) to resolve these inputs into standardized identifiers (EFO IDs, Ensembl IDs), which mitigates the risk of processing malicious instruction-based strings.
Audit Metadata