reactome-database
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill performs legitimate scientific data retrieval and analysis from a reputable source. Analysis of the scripts and instructions revealed no malicious behavior or security risks.
- [COMMAND_EXECUTION]: The skill uses
uvto run a local script. The implementation uses secure argument parsing and restricts subcommands to a pre-defined list, preventing arbitrary command injection. - [DATA_EXFILTRATION]: Network operations are restricted to the official Reactome domain (reactome.org). No unauthorized external data transmission or credential harvesting patterns were detected.
- [PROMPT_INJECTION]: The skill manages potential indirect injection risks by instructing the agent to use structured parsing tools and avoid direct ingestion of large external data sets into the prompt context.
Audit Metadata