x-brain-workspace-orientation

Pass

Audited by Gen Agent Trust Hub on Feb 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION] (LOW): The skill instructs the agent to execute a local setup script and environment management tools. \n
  • Evidence: Instructions to run ./setup.sh, poetry install, and poetry run within the ~/brain workspace. \n- [DATA_EXFILTRATION] (LOW): The skill provides explicit guidance for the agent to access sensitive configuration files and personal data. \n
  • Evidence: Mentions of ~/.ssh/config for identity management and the obsidian/ directory for personal knowledge vaults. \n- [PROMPT_INJECTION] (LOW): The skill creates a vulnerability surface for indirect prompt injection by allowing project-specific instructions to override core principles. \n
  • Ingestion points: .cursorrules, AGENTS.md, and README.md in sub-directories. \n
  • Boundary markers: Absent; the skill explicitly states that project-level rules "take precedence." \n
  • Capability inventory: Subprocess execution via poetry and ./setup.sh, file-read access to sensitive paths, and network access via referenced APIs (OpenAI, Gemini). \n
  • Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 21, 2026, 04:13 AM