GCD AI RADAR
incidents/AIR-2026-0342

OpenClaw Command and Control via Indirect Prompt Injection

Attack on AI1 Research / proof of conceptPrompt injectionAgent hijack

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.

Details

type
Attack on AI
An AI model, agent or AI-powered product was the thing exploited, abused or leaked.
categories
  • Prompt injectionprimary
    Untrusted content (web pages, documents, emails, tool output) steers a model into ignoring its instructions.
  • Agent hijack
    An AI agent is tricked into unauthorized actions: tool misuse, data exfiltration through the agent, or acting against its principal.
owasp llm top 10
mitre atlas
Both rows are derived from the type and categories (OWASP 2025 edition, ATLAS release 2026.09), not classified per record. How
severity
1 · Research / proof of concept
Demonstrated by researchers; no known real-world exploitation.
targets
Framework / tooling
products
OpenClaw
incident date
3 Feb 2026
first reported
6 Feb 2026
origin
MITRE ATLAS
Derived from MITRE ATLAS (Apache 2.0, © MITRE).
classifier confidence
0.95 · published at or above 0.80; see the methodology
tags
openclaw · prompt_injection · agent_hijack · c2
record
created 2026-10-18 06:55Z · updated 2026-10-18 06:55Z · json

Sources · 2

  1. 2026-02-06
    OpenClaw Command & Control via Prompt Injection
    MITRE ATLAS · atlas.mitre.org/studies/AML.CS0051
  2. 2026-02-06
    Exploring the Security Risks of AI Assistants like OpenClaw
    HiddenLayer · hiddenlayer.com/research/exploring-the-security-risks-of-ai-assistants-like-open

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

Related incidents