open-targets-search
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local scripts (
scripts/search) to perform search operations and uses afindcommand to resolve the script path within the plugin cache directory. This is standard behavior for resolving local script paths.\n- [EXTERNAL_DOWNLOADS]: The skill makes network requests to the Valyu API (api.valyu.ai) to perform searches. This is the primary function of the skill and uses the official provider's endpoint.\n- [CREDENTIALS_UNSAFE]: The skill manages a Valyu API key by requesting it from the user and passing it to a local setup script for storage. This follows standard local configuration patterns for API-dependent skills.\n- [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it processes and displays natural language content retrieved from the external Valyu API.\n - Ingestion points: The output of the
scripts/searchcommand, which contains data from the external Open Targets database.\n - Boundary markers: The results are presented without explicit delimiters or instructions to the agent to disregard instructions embedded in the data.\n
- Capability inventory: The skill can execute local scripts and interact with the file system via the search and setup commands.\n
- Sanitization: There is no evidence of sanitization or filtering of the external API content before it is processed by the agent.
Audit Metadata