Storia in 4 fonti

Ghostjacking: Turning Logs and Alerts into Commands to Take Over AI Agents

Ghostjacking: Turning Logs and Alerts into Commands to Take Over AI Agents 1....

Raccontata da

Confronto fonti

4 prospettive sulla stessa storia

AI · summaries

dev.toStai leggendo3 g fa

Ghostjacking: Turning Logs and Alerts into Commands to Take Over AI Agents

Attackers hide commands in monitoring logs and alerts; AI agents with write access execute them as tool calls to hijack DNS, steal cloud secrets, and run code. Tech governance must isolate read from write permissions, validate untrusted data, require approvals—sandbox containment alone is insufficient.

originale

cryptobriefing.com3 g fa

Ghostjacking attack uses poisoned logs to compromise AI agents

Tenet Security unveiled Ghostjacking at DEFCON 34: poisoning Cloudflare/Datadog logs forces AI agents to execute code with 90% success (15,000+ organizations exposed). LLMs treat all input as instruction, enabling developer access without authentication—organizations must enforce strict agent autonomy limits.

Timeline cronologica

lunedì 10 agosto 2026·securityweek.com
'Ghostjacking' Attack Uses Poisoned Logs to Turn AI Agents Bad
In a Ghostjacking attack, an AI agent executes instructions planted in the log that records a blocked request word for word.
martedì 11 agosto 2026·darkreading.com
"GhostJacking" Exposes Identity Governance Gaps in AI Agents
New research shows how attackers can use security alerts and blocked events to manipulate and hijack AI agents.

Ghostjacking: Turning Logs and Alerts into Commands to Take Over AI Agents

Ghostjacking attack uses poisoned logs to compromise AI agents

Timeline cronologica

'Ghostjacking' Attack Uses Poisoned Logs to Turn AI Agents Bad

"GhostJacking" Exposes Identity Governance Gaps in AI Agents

'Ghostjacking' Attack Uses Poisoned Logs to Turn AI Agents Bad

"GhostJacking" Exposes Identity Governance Gaps in AI Agents

Ghostjacking: Turning Logs and Alerts into Commands to Take Over AI Agents

Ghostjacking attack uses poisoned logs to compromise AI agents