NYC

adaptyv

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEDATA_EXFILTRATION
Full Analysis
  • DATA_EXFILTRATION (LOW): The skill performs network operations to a non-whitelisted domain (AWS Lambda URL). While common in serverless architectures, this constitutes an untrusted endpoint under static analysis rules. \n- EXTERNAL_DOWNLOADS (SAFE): The skill requests standard, reputable Python packages (requests, python-dotenv) for its operation. \n- PROMPT_INJECTION (SAFE): No patterns of behavior override or safety bypass were detected in the instructions. \n- INDIRECT_PROMPT_INJECTION (LOW): \n
  • Ingestion points: Processes protein sequences and experimental results retrieved from an external API. \n
  • Boundary markers: Absent in the usage examples. \n
  • Capability inventory: Utilizes network requests (requests.post) to transmit data and webhooks. \n
  • Sanitization: None demonstrated in code; the agent should treat external API responses as untrusted data.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:27 PM