deep-agents-memory

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • Filesystem Operations: The skill introduces tools for file management, including reading, writing, and editing. These capabilities are intended for local development and are supported by a 'virtual_mode' feature designed to restrict file access to a specific root directory.
  • Content Ingestion Considerations: By allowing the agent to read external file data, the skill creates a surface where the agent might encounter instructions embedded within those files. The documentation addresses this by suggesting the use of 'interrupt_on' configurations to ensure human oversight for critical file modifications.
  • Data Handling Patterns: The skill describes various backends for data persistence, ranging from ephemeral thread-based memory to persistent database storage. It provides clear guidance on selecting the appropriate backend based on the deployment environment's security requirements.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 10:12 PM