OpenClaw's Gluttony: Unbounded Memory Consumption in Webhooks
Vulnerability ID: GHSA-Q447-RJ3R-2CGH
CVSS Score: 7.5
Published: 2026-02-18
OpenClaw (formerly ClawdBot) suffers from a critical Denial of Service vulnerability due to improper handling of incoming webhook requests. The application buffers the entire request body into memory without enforcing size limits or checking the Content-Length header.
This architectural oversight allows an unauthenticated attacker to send a single, massive HTTP request—potentially gigabytes in size—forcing the Node.js process to allocate memory until it hits the V8 heap limit or triggers the OS Out-Of-Memory (OOM) killer, crashing the service instantly.
TL;DR
OpenClaw listens for webhooks but doesn't check how big the message is before trying to memorize it. Attackers can send a 5GB 'hello' message, causing the server to eat all available RAM and crash (OOM).
⚠️ Exploit Status: POC
Technical Details
- Vulnerability ID: GHSA-q447-rj3r-2cgh
- CWE: CWE-400 (Uncontrolled Resource Consumption)
- CVSS: 7.5 (High)
- Attack Vector: Network (Unauthenticated)
- Impact: Denial of Service (DoS)
- Fix Version: 2026.1.24-4
Affected Systems
- OpenClaw
- ClawdBot
- Moltbot
- Node.js Webhook Handlers
-
clawdbot: <= 2026.1.24-3 (Fixed in:
2026.1.24-4)
Exploit Details
- Internal: Simple curl command sending binary data to webhook endpoint
Mitigation Strategies
- Implement strict request body size limits in the application middleware.
- Deploy a reverse proxy (Nginx/Apache) to enforce
client_max_body_sizeat the edge. - Enforce authentication (Webhook Secrets) to validate sources before parsing bodies.
Remediation Steps:
- Stop the running OpenClaw instance.
- Update the
clawdbotpackage to version2026.1.24-4or later. - If using Docker, pull the latest
openclawimage. - Verify configuration ensures
webhookSecretis set and enforced. - Restart the service and monitor RAM usage.
References
Read the full report for GHSA-Q447-RJ3R-2CGH on our website for more details including interactive diagrams and full exploit analysis.
Top comments (0)