kegg-database

Pass

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No direct prompt injection or behavior override instructions were detected in the skill documentation or code.
  • [DATA_EXFILTRATION]: The skill facilitates data retrieval from the official KEGG API (rest.kegg.jp). No unauthorized access to local sensitive data or exfiltration attempts were found.
  • [REMOTE_CODE_EXECUTION]: There are no patterns of remote code download or execution. The logic is restricted to retrieving text or image data from the biological database.
  • [COMMAND_EXECUTION]: No shell commands or subprocess invocations are present in the provided scripts.
  • [PROMPT_INJECTION]: The skill processes biological data retrieved from an external API, creating a surface for potential indirect prompt injection. Ingestion points: all retrieval functions in scripts/kegg_api.py. Boundary markers: not explicitly defined in the instructions. Capability inventory: none (no write access or execution capabilities). Sanitization: data is returned as raw text without explicit sanitization.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 29, 2026, 02:23 AM