openclaw-status

name	openclaw-status
description	Check OpenClaw gateway health, channel connectivity, agent status, and recent cron job results
disable-model-invocation	true
allowed-tools	Bash, Read, Grep

OpenClaw Status Check

Check the health and status of the entire OpenClaw deployment.

Setup Detection

OPENCLAW_REPO=$(readlink ~/.openclaw/openclaw.json 2>/dev/null | sed 's|/.openclaw/openclaw.json||')

Detect deployment mode — multi-gateway or single-gateway:

TIER_CONFIGS=(~/.openclaw/configs/openclaw-*.json)
[[ -f "${TIER_CONFIGS[0]}" ]] && MULTI_GATEWAY=true || MULTI_GATEWAY=false

Steps — Multi-Gateway Mode

If MULTI_GATEWAY=true, check each tier independently:

1. Discover tiers and ports: For each config in ~/.openclaw/configs/openclaw-*.json:

TIER=$(basename "$cfg" | sed 's/openclaw-//;s/.json//')
PORT=$(grep -o '"port": *[0-9]*' "$cfg" | grep -o '[0-9]*')

2. Gateway health per tier:

curl -s http://localhost:$PORT/health

3. LaunchAgent status per tier:

launchctl print gui/$(id -u)/ai.openclaw.gateway.$TIER 2>&1 | head -5

4. Channel connectivity per tier — check each tier's log:

tail -50 ~/.openclaw-$TIER/gateway.log | grep -E "telegram|slack|whatsapp|starting provider|socket mode|error"

5. WhatsApp restart loop — check any tier whose config contains "whatsapp":

tail -50 ~/.openclaw-$TIER/gateway.log | grep -c "auto-restart attempt"

If count > 3, WhatsApp is in a restart loop (needs re-auth: delete ~/.openclaw/credentials/whatsapp/default/ and restart).

6. Agent count per tier: Read each tier's config and count entries in agents.list.

7. Cron job status: Read ~/.openclaw/cron/jobs.json and for each enabled job, report:

   • Job name
   • Agent ID (and which tier it belongs to)
   • Last run time and status
   • Consecutive errors (if any)
   • Next scheduled run

8. Recent errors per tier:

tail -100 ~/.openclaw-$TIER/gateway.log | grep -i "error\|fail\|crash" | tail -10

9. Present results as a per-tier summary table:
   • Each tier: name, port, running/stopped (PID), channel connectivity
   • Agents: total per tier, which have channel bindings
   • Cron: jobs with errors or that haven't run recently

Steps — Single-Gateway Mode

If MULTI_GATEWAY=false, fall back to single-gateway checks:

1. Read port from config:

PORT=$(grep -o '"port": *[0-9]*' ~/.openclaw/openclaw.json | grep -o '[0-9]*')

2. Gateway health:

curl -s http://localhost:$PORT/health

3. LaunchAgent status:

launchctl print gui/$(id -u)/ai.openclaw.gateway 2>&1 | head -5

4. Channel connectivity:

tail -50 ~/.openclaw/logs/gateway.log | grep -E "telegram|slack|whatsapp|starting provider|socket mode|error"

5. WhatsApp restart loop check:

tail -50 ~/.openclaw/logs/gateway.log | grep -c "auto-restart attempt"

6. Agent count: Read ~/.openclaw/openclaw.json and count entries in agents.list.

7. Cron job status: Read ~/.openclaw/cron/jobs.json — same as multi-gateway.

8. Recent errors:

tail -100 ~/.openclaw/logs/gateway.log | grep -i "error\|fail\|crash" | tail -10

9. Present results in a clear summary table:
   • Gateway: running/stopped (PID, uptime)
   • Channels: which are connected, which are failing
   • Agents: total count, which have channel bindings
   • Cron: jobs with errors or that haven't run recently