- 2026-03-05 - created dev mode on V2026.3.2
- 2026-03-06 - updated to mains V2026.3.7 + removed interactions with openclaw.json (now saves devMode in .env)
- An existing OpenClaw installation (this fork is based on V2026.3.2)
- Node.js 22.12+
- Python 3 (for the Hub notification server)
- Git
# 1. Stop the gateway
openclaw gateway stop
# 2. Back up the existing openclaw installation
mv /usr/lib/node_modules/openclaw /usr/lib/node_modules/openclaw.bak
# 3. Clone the fork
git clone https://github.com/bresleveloper/openclaw-dev-mode.git /opt/openclaw-dev-mode
# 4. Install dependencies
cd /opt/openclaw-dev-mode
npm install --ignore-scripts
# 5. Symlink our fork into the original location
# (required because the systemd gateway service points to /usr/lib/node_modules/openclaw/)
ln -s /opt/openclaw-dev-mode /usr/lib/node_modules/openclaw
# 6. Create a CLI wrapper
echo '#!/usr/bin/env bash' > /usr/local/bin/openclaw
echo 'set -euo pipefail' >> /usr/local/bin/openclaw
echo 'exec node /opt/openclaw-dev-mode/openclaw.mjs "$@"' >> /usr/local/bin/openclaw
chmod +x /usr/local/bin/openclaw
# 7. Enable dev mode
echo 'OPENCLAW_DEV_MODE=1' >> ~/.openclaw/.env
# 8. Start the gateway
openclaw gateway startcd /opt/openclaw-dev-mode && git pull && openclaw gateway restartopenclaw gateway stop
# Remove OPENCLAW_DEV_MODE=1 from ~/.openclaw/.env
sed -i '/OPENCLAW_DEV_MODE/d' ~/.openclaw/.env
rm /usr/lib/node_modules/openclaw
mv /usr/lib/node_modules/openclaw.bak /usr/lib/node_modules/openclaw
openclaw gateway start# Check config values are unredacted (API keys visible — means dev mode is active)
openclaw config get models.providersOpenClaw is AMAZING. And security is awesome for prod. And a hell of a buzz killer for dev/other situations.
I cloned, listed all security features (latest - V2026.3.2) and just added a simple flag to relax them:
# Add to ~/.openclaw/.env
OPENCLAW_DEV_MODE=1Because the beauty of any opensource project is that it's MINE and I am allowed to enjoy it to its full extent.
| ID | What it does |
|---|---|
| SEC-15a | Lighter safety section in system prompt |
| SEC-27 | Channel metadata treated as trusted (no "UNTRUSTED" wrapper) |
| SEC-59 | Skip messaging profile default in onboarding |
| SEC-67 | Default compaction mode (no safeguard) |
| SEC-70 | Skip browser navigation URL checks |
| SEC-71 | 50MB web fetch cap (instead of 2MB) |
| SEC-72 | Unredacted config in CLI (API keys visible) |
| SEC-78 | No control plane rate limiting |
| SEC-79 | 50MB prompt cap (instead of 2MB) |
| SEC-80 | Skip hooks token uniqueness check |
| SEC-96 | All env vars passed through to child processes |
| FIX-01 | Auto-bootstrap MEMORY.md in new workspaces |
Dev-mode includes a bundled Hub notification server that gives agents tools to send and receive notifications. It auto-starts on 127.0.0.1:10020 when dev-mode is active.
The Hub registers three agent tools: hub_notify, hub_pending, hub_done.
How it works: When a notification arrives (from a cron job, an app, or another agent), the Hub doesn't just store it — it forwards the notification through OpenClaw's /v1/chat/completions API. This means the notification enters the agent's active session as a real message, triggering a full cognitive response: the agent reasons about it using its system prompt, memory, and context, then delivers its response through the configured channel (WhatsApp, Telegram, etc.). The agent then calls hub_done to close the loop.
If you already run your own Hub server (e.g. on a different port), the auto-start will not interfere — it only spawns if port 10020 is free. You can also disable auto-start by removing or not running the bundled server.py, and point the plugin config at your existing hub instead.
If your existing hub uses a different port, update your openclaw config to match:
openclaw config set plugins.hub.port 10021See the Hub README for the full API reference, setup guide, and configuration options.