NewsSecurityWeek2026-07-23 15:09 UTC

OpenAI behebt Sicherheitslücke in ChatGPT, die Angriffe ermöglicht

Die Schwachstelle "AgentForger" erlaubte Angreifern, unsichtbare autonome AI-Agenten zu erstellen und fernzusteuern. OpenAI hat diese Sicherheitslücke in ChatGPT behoben, um Angriffe durch gefälschte interne AI-Insider zu verhindern.

AgentForger allows an attacker to create, insert and remotely control an invisible autonomous AI agent inside a victim organization. The post OpenAI Fixes ChatGPT Agent Flaw That Could Let Attackers Forge an AI Insider appeared first on SecurityWeek .

Weitere Empfehlungen

osv:CVE-2026-15628none
zhayujie chatgpt-on-wechat CowAgent Vision Tool vision.py Vision._download_to_data_url server-side request forgery
osv:MAL-2026-7556none
Malicious code in chatgptproxytmp (RubyGems)
osv:CVE-2026-14714none
zhayujie chatgpt-on-wechat CowAgent wx Endpoint common.py verify_server missing authentication
osv:CVE-2026-10214none
zhayujie chatgpt-on-wechat Bash Tool bash.py _get_safety_warning os command injection
osv:CVE-2026-7644none
ChatGPTNextWeb NextChat actions.ts addMcpServer improper authorization
CVE-2026-7644medium
A vulnerability has been found in ChatGPTNextWeb NextChat up to 2.16.1. Affected is the function addMcpServer of the file app/mcp/actions.ts. The manipulatio…
osv:CVE-2026-7643none
ChatGPTNextWeb NextChat API Endpoint Next.js cross-domain policy
CVE-2026-7643medium
A flaw has been found in ChatGPTNextWeb NextChat up to 2.16.1. This impacts an unknown function of the file Next.js of the component API Endpoint. Executing …

Weitere News-Einträge

Newstheregister-security2026-07-23
One ChatGPT link could smuggle a rogue AI agent into your company
Newscsoonline2026-07-06
Operationalizing Agentic AI: from assisted to autonomous
Newssecurityweek2026-06-29
OpenAI and Anthropic Limit New AI Models to Trump-Approved Customers During Cybersecurity Review
Newsbleepingcomputer2026-06-18
Leak confirms OpenAI is testing a ChatGPT for Science subscription
Newssecurityweek2026-06-08
OpenAI Rolling Out ChatGPT Account Security Controls
Newsthehackernews2026-06-06
New ChatGPT Lockdown Mode Limits Tools That Could Enable Data Exfiltration
Newsbleepingcomputer2026-06-02
OpenAI upgrades GPT-5.5, as it plans to retire legacy ChatGPT models
Newsbleepingcomputer2026-05-29
ChatGPT share links abused to host fake outage pages to deliver malware

Quelle: https://www.securityweek.com/openai-fixes-chatgpt-agent-flaw-that-could-let-attackers-forge-an-ai-insider/

Mehr zu ChatGPT