Guardian IDE

Guardian provides a governance layer that evaluates AI‑generated code changes before they are released. It ingests code produced by AI assistants, applies team‑defined architectural, security, and quality policies, and then produces a release decision accompanied by evidence. The system records who approved or overrode a decision, preserving an audit trail for high‑risk changes.

The tool is policy‑as‑code, stored in the repository, and operates locally through a desktop app, command‑line interface, CI/CD integration, and a VS Code extension. It supports multiple AI providers and includes a library of secret‑pattern detections. Users can configure strict, warning, or off modes for each gate, and the workflow captures context from prior sessions to inform decisions.

Targeted at small engineering teams that rely on AI assistance, Guardian aims to make AI‑augmented development auditable and compliant by combining automated enforcement with mandatory human approval.