ReasoningBank Intelligence

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE] (SAFE): No malicious behavior, obfuscation, or unauthorized access patterns were identified in the skill. All behaviors align with the stated purpose of providing an adaptive learning system for agents.
  • [DATA_EXPOSURE] (SAFE): The skill uses a local file ('.$reasoning-bank.db') for data persistence via AgentDB, which is appropriate for state management and does not involve data exfiltration.
  • [EXTERNAL_DOWNLOADS] (SAFE): The dependencies listed (agentic-flow and AgentDB) are prerequisites for the skill's environment and do not involve unauthorized remote code execution or suspicious download patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:44 PM