yichen-grok-consult

Warn

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: MEDIUMCREDENTIALS_UNSAFECOMMAND_EXECUTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill is designed to access and utilize sensitive local authentication data to interact with external services.
  • It directly accesses the user's Grok OAuth credentials stored at ~/.grok/auth.json.
  • It utilizes "X session cookies" for the xreach fallback mechanism to authenticate with X/Twitter services.
  • [COMMAND_EXECUTION]: The skill executes local binaries and command-line interfaces to perform its functions.
  • It executes the official Grok CLI binary located at ~/.grok/bin/grok or via the GROK_CONSULT_CLI environment variable.
  • It references and potentially invokes other local command-line adapters such as OpenCLI and xreach for data retrieval fallbacks.
  • [DATA_EXFILTRATION]: The skill's primary purpose is to send user-provided context and queries to external third-party services (xAI and X/Twitter). While the instructions include a boundary warning against sending passwords or private keys, the inherent nature of the tool involves transmitting data from the local environment to these external platforms.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 6, 2026, 03:19 AM
Security Audit — agent-trust-hub — yichen-grok-consult