chembl-search
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill documentation provides a shell snippet using the
findcommand to dynamically locate the search script within the agent's plugin cache directory. While this uses shell execution, it is a localized path-resolution helper and does not execute remote or untrusted code. - [PROMPT_INJECTION]: The skill processes data from the external ChEMBL database, creating a surface for indirect prompt injection where malicious instructions could theoretically be embedded in the scientific data.
- Ingestion points: Search results retrieved from the ChEMBL database via the Valyu API (SKILL.md).
- Boundary markers: None identified; search results are processed directly by the agent.
- Capability inventory: Local script execution (bash/node) and JSON processing via jq (SKILL.md).
- Sanitization: No explicit sanitization or instruction-filtering of the external content is mentioned in the documentation.
Audit Metadata