scientific-lab-data-management

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXFILTRATION]: The skill implements network communication with official scientific service APIs, such as api.dnanexus.com and Benchling tenant endpoints. These operations are necessary for the skill's functionality and do not involve exfiltrating sensitive local system data or credentials.
  • [EXTERNAL_DOWNLOADS]: Dependencies include reputable scientific software packages such as benchling-sdk, dxpy, and omero-py. These are standard tools widely used in bioinformatics and laboratory data management.
  • [PROMPT_INJECTION]: The skill processes external research data including DNA sequences, registry entries, and experimental protocols. While the provided code snippets do not explicitly show input sanitization or boundary markers, the risk is inherent to the skill's data management purpose and handled by general agent guardrails.
  • Ingestion points: DNA sequences and registry data in BenchlingClient; project file metadata in DNAnexusClient; protocol descriptions and steps in Protocols.io functions.
  • Boundary markers: None identified in the provided templates.
  • Capability inventory: Network POST/GET requests via the requests library; file upload and listing capabilities through platform-specific SDKs.
  • Sanitization: No explicit validation or sanitization is implemented in the code snippets.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 11:35 AM
Security Audit — agent-trust-hub — scientific-lab-data-management