Agent Ops Control Room

Automation yes. Autonomous no.

Agent Ops Control Room is a product-oriented React + Vite case study that shows how responsible automation should be governed before execution. It is not a bot demo. It is a local simulation of an Agent Ops safety layer with risk classification, policy checks, human approval, dry-runs and audit logging.

Live Project

https://agent-ops-control-room.vercel.app

Screenshot

• Main dashboard
• Mission analysis
• Policy explanation
• Approval center
• Audit log

See /docs/screenshots/README.md for the screenshot checklist.

Why I Built This

Many automation demos focus on making an agent take action. Real operational environments need something more careful: visibility, approval, risk handling and logs. This project demonstrates the control layer that should sit between an assistant and any sensitive action.

The Problem It Solves

Automation can help teams move faster, but sensitive actions can affect users, customers or external groups. Without governance, teams can lose control over who approved an action, why it happened and whether it should have been blocked.

Agent Ops Control Room models a safer workflow:

• Understand the operator task.
• Identify the target and possible external impact.
• Classify risk.
• Apply policy.
• Ask for human approval when needed.
• Generate a dry-run instead of taking real action.
• Write an audit trail.

Key Features

• Futuristic operations dashboard UI.
• Mission input and analysis workflow.
• Step-by-step agent pipeline.
• Safety Policy Engine with clear rules.
• Approval Center with Approve, Reject and Keep in HOLD actions.
• Live-looking audit log with timestamps, decisions, risk levels and explanations.
• Demo scenarios for safe dry-run, risky message, approval, rejection and missing approval.
• Status labels for HOLD, REVIEW, APPROVED, BLOCKED, DRY-RUN and AUDITED.
• Fully local simulation with no real messaging integrations.

Demo Flow

1. Open the app and start with the default mission: Send a scheduled WhatsApp update to the team group at 18:00.
2. Click Analyze Mission.
3. Show that HOLD mode blocks real execution because the task affects an external group.
4. Walk through the pipeline: intent, target, risk, policy, approval, dry-run and audit.
5. Use the Approval Center to approve or reject the request.
6. Click Execute simulation and show that no real message is sent.
7. Finish with the Audit Log to show operational traceability.

Technical Stack

• React
• Vite
• Tailwind CSS via @tailwindcss/vite
• Plain JavaScript policy engine
• Local browser state
• Node validation script

Architecture Overview

flowchart LR
  A["Operator mission"] --> B["Mission analysis"]
  B --> C["Risk classification"]
  C --> D["Policy engine"]
  D --> E{"Sensitive action?"}
  E -->|Yes| F["Human approval"]
  E -->|No| G["Dry-run allowed"]
  F --> H["Simulated execution only"]
  G --> H
  H --> I["Audit log"]

Loading

Main files:

• /src/App.jsx contains the dashboard UI and interaction state.
• /src/policyEngine.js contains deterministic local policy logic.
• /scripts/validate-policy.mjs validates core safety outcomes.
• /docs/portfolio-summary.md contains recruiter and portfolio copy.

Safety Model

This project is intentionally honest about what it does and does not do.

• HOLD mode blocks real execution.
• Sensitive actions require human approval.
• Dry-run is allowed.
• Real send is blocked until approved.
• Every action must be logged.
• Even after approval, execution is simulated locally.
• There are no secrets, credentials, WhatsApp automation or real messaging APIs.

All actions are simulated locally.

Recruiter-Oriented Explanation

This project is relevant for roles where operational judgment matters as much as technical execution.

IT Support: shows how to classify requests by risk and document outcomes before changes affect users.

Helpdesk / Service Desk: models approval gates, clear explanations and audit logs for support workflows.

Digital Operations: turns repetitive workflows into visible, controlled processes instead of hidden scripts.

AI Operations: demonstrates human-in-the-loop governance for AI-assisted actions.

Technical Support Specialist: shows the ability to explain policy decisions, impact and remediation steps clearly.

Junior Solutions / Implementation roles: demonstrates how to package workflow logic into a deployable interface with validation and clear user flows.

Practical capabilities demonstrated:

• Risk classification.
• Human approval.
• Audit logs.
• Change control mindset.
• Safe automation.
• Operational visibility.

What This Demonstrates For Recruiters

• I can build a polished React dashboard.
• I understand operational safety and governance.
• I can translate abstract AI risk into a practical workflow.
• I can create clear auditability and user-facing explanations.
• I can document, validate and prepare a project for deployment.
• I avoid misleading claims about automation capabilities.

How To Run Locally

cd /root/agent-ops-control-room
npm install
npm run dev

Open the local URL printed by Vite, usually:

http://localhost:5173

Run validation:

npm run validate:policy

Build for production:

npm run build

Deployment Instructions

Vercel

1. Push this project to GitHub.
2. Go to Vercel and import the repository.
3. Use the default Vite settings:
   • Framework preset: Vite
   • Build command: npm run build
   • Output directory: dist
4. Deploy.

Netlify

1. Push this project to GitHub.
2. Go to Netlify and add a new site from Git.
3. Use these build settings:
   • Build command: npm run build
   • Publish directory: dist
4. Deploy.

GitHub Publishing

cd /root/agent-ops-control-room
git init
git add .
git commit -m "Build Agent Ops Control Room case study"
git branch -M main
git remote add origin git@github.com:LambdaOpsLab/agent-ops-control-room.git
git push -u origin main

Future Improvements

• Add persisted audit logs using local storage.
• Add unit tests around the policy engine.
• Add exportable audit report snapshots.
• Add a read-only architecture diagram view inside the app.
• Add accessibility testing and keyboard flow documentation.
• Add screenshots after deployment.

Author

Created by LambdaOpsLab as part of an AI Operations portfolio focused on responsible automation, governance and human-in-the-loop workflows.