I built an open-source "approval gate" for AI agents — here's why

Описание и применение

InGravity – это легковесный шлюз утверждения для AI‑агентов, который проверяет действия перед их выполнением, блокируя рискованные операции и предоставляя аудит и ручное утверждение.

Задача

AI agents can perform high‑impact actions (payments, emails, file deletions) without a governance layer, risking unintended or malicious activity.

Результат

InGravity enables pre‑execution governance for AI agents, preventing unintended high‑risk actions while remaining lightweight and self‑hosted.

Рабочий процесс

• 1. Configure rules and risk thresholds in InGravity.
• 2. Deploy InGravity locally (self‑hosted).
• 3. In an n8n workflow, call InGravity’s /check-action endpoint before performing an action.
• 4. Receive a decision (allow/blocked) and risk score.
• 5. If blocked, the n8n node halts the workflow; if allowed, the workflow proceeds.
• 6. Review decisions in the dashboard and manually approve or block pending actions as needed.

Доказательства

AI agents are increasingly given real permissions — sending payments, emails, deleting files — often with little to no governance layer in between.

Источник

InGravity evaluates an action against rules you configure, and returns a decision — before the action runs.

Источник

A rules engine + risk scoring API (POST /check-action)

Источник

No authentication on the API yet (don't expose it to untrusted networks — self-host only for now)

Источник