What is WarRoom?

OpenClaw WarRoom is a high-signal TUI for supervising autonomous agent execution.

Instead of reading raw logs, you get:

• a CENTER mission board
• intelligently named task rooms
• per-task milestone progression
• nested CODING rooms with live implementation stream
• SLA-style urgency cues for fast intervention

Built for founders/operators running many parallel agent tasks.

Why it feels different

Most dashboards are either too shallow (just statuses) or too noisy (raw logs everywhere). WarRoom gives a layered model:

1. CENTER → global awareness (what matters now)
2. TASK room → intent + milestones
3. CODING room → what the agent is actually doing line-by-line

This keeps you in control without context-switch fatigue.

Core features

🎯 CENTER room

• Running / Blocked / Done / SLA summary
• Priority-oriented list of active tasks
• Pin support for mission-critical threads

🧠 Intelligent task naming

Automatic naming from real intent and commands, e.g.:

• AETHER: Train Conversations
• AETHER: Curate Memory Systems
• X: Publish / Engage
• Infra: AWS Planning
• Comms: Matrix Synapse

🪜 Milestone timeline

Every task builds a compact timeline of:

• request received
• tool execution
• progress updates
• completion / error events

🧩 Split view (milestones + live feed)

Task and coding rooms render as:

• Left pane: milestones + metadata
• Right pane: live coding / execution feed
• Smooth horizontal pan (h / l) for long lines

🔔 PINGS room (scheduled reminders tracker)

Dedicated room for scheduled "ping/remind me" workflows:

• tracks OpenClaw cron jobs (next run + status)
• shows reminder text / intent
• supports manual watch notes from any selected room (g)

⚙️ Customization menu

Open CONFIG room (m) to control visible windows and layout. Window toggles:

• 1 CENTER
• 2 PINGS
• 3 TASKS
• 4 CODING
• 5 TRACE

🛰 Nested ↳ CODING room

Each task has a child implementation room with real-time stream of:

• tool calls
• coding updates
• execution events

🚨 Operator controls

• SLA warning and alert thresholds
• Pin/unpin task rooms
• Alert-only mode
• Seamless refresh strategy to reduce flicker/reload jitter

Install

npm install -g @sendroon/openclaw-warroom
openclaw-warroom

Requires python3 (UI core is Textual).

Usage (local dev)

cd ~/.openclaw/workspace/tools/warroom-cli
./run-warroom.sh

Multi-terminal linking + custom layouts

Use same link id to share state between terminals, and different profiles for different window layouts:

# Main operator terminal
./run-warroom.sh --profile ops --link team1

# Secondary terminal: LIVE FEED mirror of selected room from main
./run-warroom.sh --profile live --link team1 --live-feed

Equivalent npm command:

openclaw-warroom --profile live --link team1 --live-feed

In LIVE FEED mode, whatever room you select in the main linked terminal is shown full-screen in the live feed terminal.

Keybindings

• mouse click select room
• j / k next / previous room
• h / l horizontal pan left/right
• p pin / unpin selected task
• g add ping entry from selected room
• m open customization menu (CONFIG room)
• 1..5 toggle windows (center, pings, tasks, coding, trace)
• a toggle alert mode
• r reload
• q quit

LIVE FEED mode keys:

• r refresh
• h / l horizontal pan
• q quit

Data sources

• Session transcripts: ~/.openclaw/agents/main/sessions/*.jsonl
• Gateway logs: ~/.openclaw/logs/*

Bug reporting

Use GitHub Issues (templates included):

• 🐞 Report bug: https://github.com/baiehclaca/openclaw-warroom/issues/new?template=bug_report.yml
• 💡 Feature request: https://github.com/baiehclaca/openclaw-warroom/issues/new?template=feature_request.yml

Maintainer notifications

To get instant notifications for new reports:

1. Open repo on GitHub → Watch
2. Choose Custom
3. Enable Issues (and optionally Pull requests)

Positioning

OpenClaw WarRoom is not a generic log tailer. It is an operator UX layer for real-time human-in-the-loop control of autonomous execution.

If OpenClaw is your engine, WarRoom is your cockpit.