gcdRadar
closedclosed 20 Feb 2026Breaches & attacks3 Confirmed exploitationAI

Exposed ClawdBot Control Interfaces Lead to Credential Access

OpenClaw, ClawdBot

A researcher discovered hundreds of unauthenticated ClawdBot control interfaces exposed on the public internet. The exposure allowed unauthorized access to configuration files, credentials, and execution of assistant skills.

first report
6 Feb 2026
last update
6 Feb 2026

Timeline · 4

  1. 6 Feb 2026Exposed ClawdBot Control Interfaces Leads to Credential Access and Execution
  2. 6 Feb 2026GitHub - openclaw/openclaw: Your own personal AI assistant. Any OS. Any Platform. The lobster way. GitHub
  3. 6 Feb 2026Clawdbot Control - Shodan Search
  4. 6 Feb 2026hacking clawdbot and eating lobster souls

Sources · 4

News and write-ups

  1. 2026-02-06
  2. 2026-02-06
    Clawdbot Control - Shodan Search
    shodan.io · shodan.io/search?query=Clawdbot+Control
  3. 2026-02-06
    hacking clawdbot and eating lobster souls
    x.com · x.com/theonejvo/status/2015401219746128322

Research and catalogues

  1. 2026-02-06

Titles and links belong to their publishers. The summary above is written by gcdRadar in its own words; data is CC BY-SA 4.0.

Classification

AI role
Attack on AI
categories
Data leakAgent hijack
severity
3 · Confirmed exploitation
targets
Enterprise deployment, Open-weight model
dates
happened 25 Jan 2026 · first reported 6 Feb 2026
origin
MITRE ATLAS
Derived from MITRE ATLAS (Apache 2.0, © MITRE).
record
confidence 0.95 · created 2026-10-18 06:55Z · updated 2026-10-18 06:55Z

Lane, AI role, categories and severity come from the classifier; the OWASP and ATLAS rows are derived from the categories. How these are assigned

Copy a brief

This record as a short Markdown brief, for a ticket, a chat or a coding agent.

the brief as text
# Exposed ClawdBot Control Interfaces Lead to Credential Access

https://gcdradar.vercel.app/incidents/exposed-clawdbot-control-interfaces-lead-to-credential-acces · AIR-2026-0343 · closed 2026-02-20

- Lane: Breaches & attacks
- Severity: 3 (Confirmed exploitation)
- Fix status: No fix needed
- First reported: 2026-02-06 (happened 2026-01-25)
- AI role: Attack on AI
- Products: OpenClaw, ClawdBot

## Summary

A researcher discovered hundreds of unauthenticated ClawdBot control interfaces exposed on the public internet. The exposure allowed unauthorized access to configuration files, credentials, and execution of assistant skills.

## What to do

No action recorded; read the sources below and the vendor's advisory.

## Latest updates

- 2026-02-06: hacking clawdbot and eating lobster souls (https://x.com/theonejvo/status/2015401219746128322)
- 2026-02-06: Clawdbot Control - Shodan Search (https://shodan.io/search?query=Clawdbot+Control)
- 2026-02-06: GitHub - openclaw/openclaw: Your own personal AI assistant. Any OS. Any Platform. The lobster way. GitHub (https://github.com/openclaw/openclaw)
- 2026-02-06: Exposed ClawdBot Control Interfaces Leads to Credential Access and Execution (https://atlas.mitre.org/studies/AML.CS0048)

## Sources

- 2026-02-06 MITRE ATLAS: Exposed ClawdBot Control Interfaces Leads to Credential Access and Execution https://atlas.mitre.org/studies/AML.CS0048
- 2026-02-06 GitHub: GitHub - openclaw/openclaw: Your own personal AI assistant. Any OS. Any Platform. The lobster way. GitHub https://github.com/openclaw/openclaw
- 2026-02-06 shodan.io: Clawdbot Control - Shodan Search https://shodan.io/search?query=Clawdbot+Control
- 2026-02-06 x.com: hacking clawdbot and eating lobster souls https://x.com/theonejvo/status/2015401219746128322

Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.

Related

  1. closed
    AI
  2. closed
    Hermes Suspected Hackers Used AI Agents to Compromise Taiwanese Systems

    Breaches & attacks · Real loss · since 1 Jul 2026

    AI
  3. closed
    AI
  4. closed
    AI
  5. closed
    AI
  6. closed
    AI