deep-agents-core

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFE
Full Analysis
  • Framework Configuration: The skill provides templates for setting up agents with specific capabilities. It includes security-conscious recommendations, such as using interrupts for file-writing operations to ensure human oversight.
  • Filesystem Tools: The documentation describes tools for reading, writing, and editing files. While these are powerful capabilities, the skill demonstrates how to configure these with backends (e.g., FilesystemBackend) and recommends restricting access to a specific root directory.
  • Task Delegation: The framework supports spawning subagents for specialized tasks. The documentation notes that skills are not automatically inherited by subagents, requiring explicit configuration, which follows the principle of least privilege.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 11:48 AM
Security Audit — agent-trust-hub — deep-agents-core