kegg-database
Pass
Audited by Gen Agent Trust Hub on Mar 31, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches data from the well-known KEGG REST API (rest.kegg.jp). These operations are performed using standard Python libraries for scientific data retrieval and do not include the execution of remote scripts.
- [SAFE]: No evidence of malicious code, prompt injection instructions, or unauthorized credential access was found in the instructions or scripts. The helper functions are focused on API interaction with no local system modification capabilities.
- [PROMPT_INJECTION]: Indirect prompt injection surface analysis:
- Ingestion points: External data enters the agent context via API response text in
scripts/kegg_api.py(functions:kegg_info,kegg_list,kegg_find,kegg_get,kegg_conv,kegg_link,kegg_ddi). - Boundary markers: Absent. External content is not delimited by markers or safety warnings.
- Capability inventory: Restricted to network read operations (
urllib.request.urlopen) inscripts/kegg_api.py. No dangerous tools like shell execution or file system writes are available to the skill. - Sanitization: Absent for retrieved data; however, search queries are URL-encoded for API path safety. The risk is assessed as safe given the curated academic nature of the data source.
Audit Metadata