search-brain

Pass

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The search query provided in $ARGUMENTS is directly interpolated into the pattern parameters for the Grep and Glob tools. This creates a surface where a user could attempt to craft a query that manipulates the tool's intended behavior or parameters.\n- [PROMPT_INJECTION]: The skill ingests untrusted content from the 'Brain/' directory, creating an indirect prompt injection vulnerability.\n
  • Ingestion points: The skill reads file names and content from the 'Brain/' directory using Grep, Glob, and Read tools.\n
  • Boundary markers: No delimiters or instructions are used to distinguish search results from agent instructions, which could lead the agent to follow directives embedded in notes.\n
  • Capability inventory: The skill utilizes tools for filesystem access, specifically Read, Grep, and Glob.\n
  • Sanitization: Note content is ranked and presented without any sanitization or validation to filter out potential malicious instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 5, 2026, 11:46 PM