target-drug-evidence-evaluation
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data retrieved from pharmaceutical databases, clinical trial reports, and patent registries.
- Ingestion points: Data enters the context through tools such as
ls_clinical_trial_result_search,ls_clinical_guideline_vector_search, andls_patent_fetch(specifically for analyzing patent claims). - Boundary markers: Absent. There are no specific delimiters or instructions provided to the agent to disregard potential instructions embedded within the retrieved technical or legal text.
- Capability inventory: The agent's capabilities are constrained to a predefined set of analytical pharmaceutical and patent search tools. It does not possess capabilities for arbitrary file system access, network exfiltration to unknown domains, or shell command execution.
- Sanitization: None detected. The skill relies on the structured nature of the pharmaceutical data provided by the underlying MCP tools.
Audit Metadata