metabolomics-workbench-database
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill facilitates access to the NIH Metabolomics Workbench through its official REST API. All network communications are directed to the service's official domain at metabolomicsworkbench.org. The examples provided use standard Python requests and curl commands to fetch scientific data and structures. No obfuscation, credential exposure, or persistence mechanisms were detected.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from an external API, which theoretically presents a surface for indirect prompt injection if the remote data contains malicious instructions designed to be interpreted by an agent. Ingestion points: API responses from https://www.metabolomicsworkbench.org/rest/ are processed by the agent (identified in SKILL.md and references/api_reference.md). Boundary markers: None identified in the provided instructions to explicitly ignore instructions within data. Capability inventory: The skill demonstrates capabilities for network requests (requests.get) and file writing (saving structure images). Sanitization: No explicit sanitization of API responses is documented in the usage examples.
Audit Metadata