open-targets-search
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill employs a Bash script wrapper to execute its Node.js logic, which is a standard pattern for providing a CLI interface.
- [DATA_EXFILTRATION]: User-provided queries are transmitted to the Valyu API at
api.valyu.aito retrieve drug-disease association data. This is the intended purpose of the skill. - [INDIRECT_PROMPT_INJECTION]: The skill processes and returns external search results, which constitutes an attack surface for indirect prompt injection. 1. Ingestion points: Data is ingested from the Valyu API via
fetchinscripts/search.mjs. 2. Boundary markers: The skill outputs results in a structured JSON format, providing a clear boundary between data and agent instructions. 3. Capability inventory: The skill's capabilities are limited to making network requests to a specific domain and managing its own configuration file. 4. Sanitization: No explicit sanitization of the retrieved content is performed before it is returned to the agent.
Audit Metadata