governance-gate

What problem does it solve?

Governance Gate enforces safety, policy, and recovery controls across autonomous agent executions to prevent destructive, privacy-leaking, or non-compliant operations and to ensure actions are auditable and reversible.

Core Features & Use Cases

• Five safety layers: input sanitization, policy gating, execution sandbox, output validation, and recovery/rollback.
• Seven policy constraints: safety, privacy, data access, financial, compliance, fairness, and transparency checks that block, redact, warn, or escalate as needed.
• Escalation tiers and capability gating: T0–T3 risk tiers with explicit escalation behavior and enforcement of MCP server tool bindings; side-effect logging and git-safe workflows support recovery.
• Use Cases: validating CI/CD deploys, gating database migrations, preventing secret leakage during code edits, and escalating financial or trading operations for human approval.

Quick Start

Ask governance-gate to evaluate and enforce policies for a proposed agent plan that would push infrastructure changes to production and return a compliant, escalated execution plan.