loaditout.ai
SkillsPacksTrendingLeaderboardAPI DocsBlogSubmitRequestsCompareAgentsXPrivacyDisclaimer
{}loaditout.ai
Skills & MCPPacksBlog

nornr-sentry

MCP Tool

NORNR/nornr-sentry

NORNR Sentry local airbag for dangerous agent actions

Install

$ npx loaditout add NORNR/nornr-sentry

Platform-specific configuration:

.claude/settings.json
{
  "mcpServers": {
    "nornr-sentry": {
      "command": "npx",
      "args": [
        "-y",
        "nornr-sentry"
      ]
    }
  }
}

Add the config above to .claude/settings.json under the mcpServers key.

About

NORNR Sentry

NORNR Sentry is a local decision layer for consequential agent actions.

This public repo is the open wedge:

  • one dangerous action
  • one mandate conflict
  • one human choice
  • one defended record afterward

It is not the hosted NORNR control plane.

Operator station

After install, Sentry opens into a local operator station for patch / wiring, verify, replay, records, proof hub and serve flows. Use the blocked stop-screen as the first proof image, and this screen as the second image that shows the product is a real navigable tool after the first stop.

Proof
  • Hero stop-screen: nornr-sentry-blocked-stop-screen.png
  • Operator station: nornr-sentry-operator-station.png
  • Proof clip: nornr-sentry-proof-clip-final.mp4
  • X-optimized clip: nornr-sentry-proof-clip-x.mp4
Install

Open the app:

npx nornr-sentry

Open the direct install chooser:

npx nornr-sentry --patch-client

Or install globally:

npm install -g nornr-sentry
Public proof flow
  1. Open Sentry or choose patch / wiring.
  2. Verify the real target.
  3. Run one demo stop.
  4. Observe first in shadow mode.
  5. Serve for real.
  6. Open the proof queue, export the defended record, and replay it locally.

Or clone and run locally:

npm install
npm run demo:cursor
NPM release
npm run qa:public-package
cd ../../dist/nornr-sentry-public
npm publish
What is in this public repo
  • local proxy runtime
  • local TUI review
  • patch flow for Cursor and Claude Desktop
  • local mandate init and tighten loop
  • policy replay demo
  • shadow mode and shadow conversion

Tags

ai-agentsai-safetyclaude-desktopcursormcppolicy-gatetui

Reviews

Loading reviews...

Quality Signals

0
Installs
Last updated16 days ago
Security: AREADME

Safety

Risk Levelmedium
Data Access
read
Network Accessnone

Details

Sourcegithub-crawl
Last commit3/30/2026
View on GitHub→

Embed Badge

[![Loaditout](https://loaditout.ai/api/badge/NORNR/nornr-sentry)](https://loaditout.ai/skills/NORNR/nornr-sentry)