drug-discovery-search
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION] (SAFE): The skill documentation instructs the agent to locate and execute local bash and Node.js scripts. This is the intended architecture for the tool and no privilege escalation or dangerous parameters are suggested.
- [DATA_EXFILTRATION] (SAFE): Network communication is directed solely to the documented API provider at api.valyu.ai for searching biological data.
- [PROMPT_INJECTION] (LOW): The skill is susceptible to indirect prompt injection due to its ingestion of untrusted natural language data. 1. Ingestion points: The user-provided query argument passed to the search script. 2. Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the documentation. 3. Capability inventory: The skill has network access to perform API requests and processes returned data. 4. Sanitization: No sanitization logic is described in the integration documentation.
Audit Metadata