Executive Summary
Ghostjacking makes an AI agent execute instructions an attacker planted in a log or alert that records a blocked request word for word.
What Happened
An AI agent executes instructions that an attacker has planted in the log or alert that records a blocked request word for word. The post ‘Ghostjacking’ Attack Uses Poisoned Logs to Turn AI Agents Bad appeared first on SecurityWeek .
Vulnerability / Attack Technique
Prompt injection via logs of blocked requests
Technical Mechanism
The attacker sends a request containing instructions; it is blocked but logged verbatim, and the agent that later reads the log follows the instructions.
AI Security Category
LLM01:2025 Prompt Injection
Part of the Prompt Injection topic.
Relevant Frameworks
RealGround Analysis
Blocking the request is not enough if the log of it becomes the attack. RealGround analysis: AI agents reading security telemetry must isolate attacker-controlled fields.
Recommended Controls
- Mark or escape attacker-controlled fields in logs before agents read them
- Limit what log-triage agents can do
Related RealGround Service
AI Agent Business Logic Audit · or start with the free AI Agent Security Scanner.
Sources
securityweek.com: https://www.securityweek.com/ghostjacking-attack-uses-poisoned-logs-to-turn-ai-agents-bad/
Summary and analysis written by RealGround from the source above; last updated 2026-10-18.