brain-expert

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMCOMMAND_EXECUTION
Full Analysis
  • Persistence Mechanisms (MEDIUM): The skill includes instructions for the 'brain cron install' command, which modifies the system's crontab to establish a persistent background synchronization task. Granting an agent the ability to create persistent tasks is a high-privilege operation.
  • Indirect Prompt Injection (LOW): The agent is tasked with searching and displaying content from user-controlled Markdown files. Maliciously crafted notes could contain instructions aimed at manipulating the agent's behavior. \n
  • Ingestion points: The 'brain search' and 'brain show' commands read content from the local knowledge base. \n
  • Boundary markers: The skill does not provide delimiters or instructions for the agent to ignore potential commands within the notes. \n
  • Capability inventory: The agent has the ability to delete files ('brain rm'), modify files ('brain add'), and install background tasks ('brain cron'). \n
  • Sanitization: There is no mention of content sanitization or validation for the Markdown files processed.
  • Command Execution (LOW): The skill provides a comprehensive interface to the 'brain' CLI, allowing the agent to perform extensive file system operations, including creating, listing, and deleting files within the knowledge base directory.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:16 PM