tooluniverse-metabolomics-pathway
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill explicitly instructs the agent to write and run Python code via Bash to handle computation, statistics, and data processing. This involves the runtime generation and execution of code within the agent's environment.
- [COMMAND_EXECUTION]: The instructions require the use of Bash as a wrapper to execute Python scripts for data analysis.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from multiple external biological databases (KEGG, Reactome, MetaCyc, HMDB, CTD) and processes it using script execution capabilities. This creates a vulnerability surface if the external data source were to contain malicious instructions targeting the agent's processing logic.
- Ingestion points: External data retrieved via tools such as
Metabolite_search,KEGG_get_compound, andCTD_get_chemical_diseasesinSKILL.md. - Boundary markers: The instructions do not specify any delimiters or safety warnings to ignore instructions embedded in the retrieved data.
- Capability inventory: The skill has access to Bash and Python for data analysis (defined in the 'COMPUTE, DON'T DESCRIBE' section).
- Sanitization: There are no instructions for sanitizing or validating the structure of the data retrieved from external tools before interpolation into processing scripts.
Audit Metadata