{
 "id": "AIR-2026-0342",
 "slug": "openclaw-command-and-control-via-indirect-prompt-injection",
 "title": "OpenClaw Command and Control via Indirect Prompt Injection",
 "summary": "Researchers demonstrated how a malicious webpage embedding an indirect prompt injection could cause OpenClaw to silently execute scripts and plant persistent instructions. The technique turned the assistant into an unauthorized command and control agent.",
 "type": "attack_on_ai",
 "categories": [
  "prompt_injection",
  "agent_hijack"
 ],
 "severity": 1,
 "targets": [
  "framework_or_tooling"
 ],
 "products": [
  "OpenClaw"
 ],
 "incident_date": "2026-02-03",
 "incident_date_precision": "day",
 "first_reported": "2026-02-06",
 "sources": [
  {
   "url": "https://atlas.mitre.org/studies/AML.CS0051",
   "publisher": "MITRE ATLAS",
   "title": "OpenClaw Command & Control via Prompt Injection",
   "published_at": "2026-02-06"
  },
  {
   "url": "https://hiddenlayer.com/research/exploring-the-security-risks-of-ai-assistants-like-openclaw",
   "publisher": "HiddenLayer",
   "title": "Exploring the Security Risks of AI Assistants like OpenClaw",
   "published_at": "2026-02-06"
  }
 ],
 "source_count": 2,
 "confidence": 0.95,
 "status": "published",
 "origin": "seed_atlas",
 "tags": [
  "openclaw",
  "prompt_injection",
  "agent_hijack",
  "c2"
 ],
 "created_at": "2026-10-18T06:55:08.239Z",
 "updated_at": "2026-10-18T06:55:08.239Z"
}