deep-agents-core

Pass

Audited by Gen Agent Trust Hub on Mar 4, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • Filesystem Interaction Capabilities: The documentation describes built-in tools for local environment interaction, including read_file, write_file, and edit_file. These capabilities are presented as core features of the framework's context management middleware.
  • Agent Orchestration and Delegation: The skill introduces a task tool used for spawning specialized subagents. This facilitates multi-agent task decomposition and is a standard architectural pattern for complex agentic workflows.
  • Data Ingestion Surface: The framework is designed to process external content through its filesystem backend and dynamic skill loading. While this represents a surface for processing untrusted data, the skill focuses on the structural implementation of these features within a secure development lifecycle.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 4, 2026, 03:32 PM