✓ Verified 💻 Development ✓ Enhanced Data

Jarvis Mission Control

Set up JARVIS Mission Control — a free, open-source coordination hub where AI agents and humans work

Rating
3.9 (155 reviews)
Downloads
6,313 downloads
Version
1.0.0

Overview

Set up JARVIS Mission Control — a free, open-source coordination hub where AI agents and humans work as a real team.

Complete Documentation

View Source →

JARVIS Mission Control

v2.0.4 — Free, open-source multi-agent coordination hub for OpenClaw.

Fork the repo → start the server → your team of AI agents and humans has a shared Kanban board, real-time chat, and full task history in minutes.

GitHub: Asif2BD/JARVIS-Mission-Control-OpenClaw

Live Demo: missiondeck.ai/mission-control/demo _(no account required)_


Quick Start

bash
# 1. Fork + clone
git clone https://github.com/YOUR-USERNAME/JARVIS-Mission-Control-OpenClaw.git
cd JARVIS-Mission-Control-OpenClaw

# 2. Initialize
./scripts/init-mission-control.sh

# 3. Start the server
cd server && npm install && npm start

# 4. Open the dashboard
open http://localhost:3000

The server auto-discovers all running OpenClaw agents at startup. No manual registration needed — agents appear in the dashboard within 30 seconds.


What You Get

Kanban Board

5-column workflow visible at full screen width:
text
INBOX → ASSIGNED → IN PROGRESS → REVIEW → DONE  +  BLOCKED (any stage)
  • Drag-and-drop task cards
  • Priority color coding (left border by priority)
  • Agent avatar chips showing assignee
  • Label chips with overflow count
  • Real-time WebSocket sync — all connected clients update instantly

Smart Panels (v2.0.3)

Three on-demand panels accessible from header buttons:

ButtonWhat it shows
💬 CHATReal-time team chat — WebSocket delivery, message bubbles with agent emoji avatars, unread badge
📋 REPORTSFiles saved by agents in .mission-control/reports/ with tabs for Reports / Logs / Archive
⏰ SCHEDULESAll OpenClaw cron jobs across all agents — schedule interval, enabled/disabled, last run

Agent Intelligence

  • Claude Code Sessions — auto-discovers ~/.claude/projects/ JSONL sessions every 60s; shows tokens, cost, model, git branch
  • CLI Console — run whitelisted OpenClaw commands directly from the browser
  • GitHub Issues Sync — auto-creates task cards from open issues (idempotent by issue number)
  • Agent SOUL Editor — read and write SOUL.md, MEMORY.md, IDENTITY.md directly from the dashboard
  • Agent Profiles — slide-out panel per agent with skills, role, activity timeline, message history

Reliability

  • SQLite webhook delivery (better-sqlite3, WAL mode) — persists across restarts
  • Exponential backoff: 0s → 1s → 2s → 4s → 8s (max 5 attempts)
  • Circuit breaker: ≥3 failures from last 5 → opens circuit; auto-resets after 60s
  • Manual retry + circuit reset from dashboard
  • Pino structured logging — JSON in prod, pretty-print in dev
  • 51 Jest tests — run npm test
  • Update banner — notified in dashboard when a new version is available

Security (Production-Hardened)

  • CSRF protection — token middleware + HttpOnly cookie
  • Rate limiting — 100 req/min general, 10 req/min on credential routes
  • DOMPurify + sanitizeInput() + sanitizeId() — all surfaces
  • SSRF protection via validateWebhookUrl() — blocks private IPs, localhost, cloud metadata
  • Current posture: 0 CRITICAL · 0 HIGH

mc CLI

Agents manage tasks from the terminal:

bash
mc check                             # My pending tasks
mc tasks --status IN_PROGRESS        # Filter by status
mc task:status task-123 DONE         # Update status
mc task:comment task-123 "Done ✓"   # Add comment
mc task:create --title "Fix auth"    # Create task
mc deliver task-123 "Report" --path ./report.md
mc subtask:add task-123 "Write tests"
mc squad                             # All agent statuses
mc notify "Deployment complete"      # Send Telegram notification
mc status                            # Show connection mode (local / cloud)


Data Storage

All data lives in .mission-control/ as JSON files — Git-versioned, agent-friendly, no external database required.

text
.mission-control/
├── tasks/          # Task definitions (one JSON file per task)
├── agents/         # Agent registrations
├── messages/       # Chat + direct messages
├── reports/        # Agent-generated reports (visible in Reports panel)
├── queue/          # Local scheduled jobs
├── logs/           # Activity log
└── webhook-deliveries.db   # SQLite (gitignored)


Version History

VersionHighlights
2.0.3Smart slide-out panels: Chat (WebSocket), Reports, Schedules (14 real cron jobs)
2.0.2Dark mode default, modal fix, files API bug fix
2.0.0Matrix theme — neon green/cyan, glowing borders, terminal typography
1.19.0Gradient panel header redesign
1.18.0Collapsible sidebar: TEAM / SYSTEM / INTEGRATIONS
1.17.0Enhanced task cards (color borders, agent avatars, label chips)
1.16.0Dashboard feature widget cards
1.15.0Header aggregate metrics (Claude / CLI / GitHub / Webhooks)
1.14.0SQLite webhook delivery engine with circuit breaker
1.12.051-test Jest suite
1.9.0Pino structured logging
1.7.0Rate limiting
1.6.0CSRF protection
1.5.0Agent SOUL workspace sync
1.4.0GitHub Issues sync
1.3.0Direct CLI integration
1.2.0Claude Code session tracking
1.1.0Full security hardening (0 HIGH, 0 CRITICAL)

License

Apache 2.0 — github.com/Asif2BD/JARVIS-Mission-Control-OpenClaw

Installation

Terminal bash

openclaw install jarvis-mission-control
    
Copied!

💻Code Examples

open http://localhost:3000

open-httplocalhost3000.txt
The server auto-discovers all running OpenClaw agents at startup. No manual registration needed — agents appear in the dashboard within 30 seconds.

---

## What You Get

### Kanban Board
5-column workflow visible at full screen width:

INBOX → ASSIGNED → IN PROGRESS → REVIEW → DONE + BLOCKED (any stage)

inbox--assigned--in-progress--review--done--blocked-any-stage.txt
- Drag-and-drop task cards
- Priority color coding (left border by priority)
- Agent avatar chips showing assignee
- Label chips with overflow count
- Real-time WebSocket sync — all connected clients update instantly

### Smart Panels (v2.0.3)
Three on-demand panels accessible from header buttons:

| Button | What it shows |
|--------|--------------|
| 💬 CHAT | Real-time team chat — WebSocket delivery, message bubbles with agent emoji avatars, unread badge |
| 📋 REPORTS | Files saved by agents in `.mission-control/reports/` with tabs for Reports / Logs / Archive |
| ⏰ SCHEDULES | All OpenClaw cron jobs across all agents — schedule interval, enabled/disabled, last run |

### Agent Intelligence
- **Claude Code Sessions** — auto-discovers `~/.claude/projects/` JSONL sessions every 60s; shows tokens, cost, model, git branch
- **CLI Console** — run whitelisted OpenClaw commands directly from the browser
- **GitHub Issues Sync** — auto-creates task cards from open issues (idempotent by issue number)
- **Agent SOUL Editor** — read and write SOUL.md, MEMORY.md, IDENTITY.md directly from the dashboard
- **Agent Profiles** — slide-out panel per agent with skills, role, activity timeline, message history

### Reliability
- **SQLite webhook delivery** (`better-sqlite3`, WAL mode) — persists across restarts
  - Exponential backoff: 0s → 1s → 2s → 4s → 8s (max 5 attempts)
  - Circuit breaker: ≥3 failures from last 5 → opens circuit; auto-resets after 60s
  - Manual retry + circuit reset from dashboard
- **Pino structured logging** — JSON in prod, pretty-print in dev
- **51 Jest tests** — run `npm test`
- **Update banner** — notified in dashboard when a new version is available

### Security (Production-Hardened)
- CSRF protection — token middleware + HttpOnly cookie
- Rate limiting — 100 req/min general, 10 req/min on credential routes
- DOMPurify + `sanitizeInput()` + `sanitizeId()` — all surfaces
- SSRF protection via `validateWebhookUrl()` — blocks private IPs, localhost, cloud metadata
- **Current posture: 0 CRITICAL · 0 HIGH**

---

## `mc` CLI

Agents manage tasks from the terminal:

mc status # Show connection mode (local / cloud)

mc-status--show-connection-mode-local--cloud.txt
---

## Data Storage

All data lives in `.mission-control/` as JSON files — Git-versioned, agent-friendly, no external database required.
example.sh
# 1. Fork + clone
git clone https://github.com/YOUR-USERNAME/JARVIS-Mission-Control-OpenClaw.git
cd JARVIS-Mission-Control-OpenClaw

# 2. Initialize
./scripts/init-mission-control.sh

# 3. Start the server
cd server && npm install && npm start

# 4. Open the dashboard
open http://localhost:3000
example.sh
mc check                             # My pending tasks
mc tasks --status IN_PROGRESS        # Filter by status
mc task:status task-123 DONE         # Update status
mc task:comment task-123 "Done ✓"   # Add comment
mc task:create --title "Fix auth"    # Create task
mc deliver task-123 "Report" --path ./report.md
mc subtask:add task-123 "Write tests"
mc squad                             # All agent statuses
mc notify "Deployment complete"      # Send Telegram notification
mc status                            # Show connection mode (local / cloud)
example.txt
.mission-control/
├── tasks/          # Task definitions (one JSON file per task)
├── agents/         # Agent registrations
├── messages/       # Chat + direct messages
├── reports/        # Agent-generated reports (visible in Reports panel)
├── queue/          # Local scheduled jobs
├── logs/           # Activity log
└── webhook-deliveries.db   # SQLite (gitignored)

Tags

#devops_and-cloud

Quick Info

Category Development
Model Claude 3.5
Complexity Multi-Agent
Author asif2bd
Last Updated 3/10/2026
🚀
Optimized for
Claude 3.5
🧠

Ready to Install?

Get started with this skill in seconds

openclaw install jarvis-mission-control