Directory
‹ Back to apps

malanta-trustgate

Checks where your agent is about to send bytes - domains, IPs, and GitHub repositories — against a reputation provider before the action runs, and blocks what comes back flagged. Covers shell commands, MCP calls, file reads, and WebFetch/WebSearch. Malanta by default, or bring your own vendor.

What’s inside

1 skill
  • SkilltrustgateOperate and troubleshoot Malanta TrustGate — the Cursor security hooks that check domains/IPs (reputation) and behavioral attack patterns (ATR) before shell, MCP, file-read, and WebFetch/WebSearch actions. Use when a TrustGate hook blocks, warns, or asks for human approval on an action, when interpreting or explaining a verdict, when a domain or command was denied/allowed/paused for approval, or when configuring TrustGate's mode, scope, overrides, retention, or provider.