unmet-clinical-need-extractor
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No evidence of malicious behavior, credential exfiltration, or unauthorized command execution was found. The skill consists entirely of natural language instructions and reference frameworks for biomedical analysis. References to external resources point to the author's own GitHub repository (aipoch), which is considered a standard vendor resource.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external biomedical literature, which represents a surface for potential indirect prompt injection attacks. 1. Ingestion points: Step 2 in the execution process retrieves guidelines, reviews, and clinical studies from external sources. 2. Boundary markers: The skill does not explicitly define delimiters to separate system instructions from untrusted data, though it does mandate strict adherence to retrieved evidence. 3. Capability inventory: The skill does not utilize any high-risk capabilities such as arbitrary shell command execution, local file system writing, or persistence mechanisms. 4. Sanitization: There are no specific instructions for sanitizing ingested data for malicious injection patterns, however, the risk is mitigated by the analytical nature of the task and lack of sensitive tool access.
Audit Metadata