Become a Member

Regulation TrackerLaws, orders and treaties read for their security duties and clocks Standards & FrameworksAttack taxonomies, control catalogues and frontier frameworks AI Incident DatabaseThe security incident record, with mechanism and lesson Research Paper LibraryThe papers that define the field, summarised OSS DirectoryRed teaming, guardrail, supply chain and evaluation tools Case StudiesProgrammes built, with measured outcomes AI Vulnerability TrackerCVEs and advisories for the components AI systems are built from

Latest in the tracker

New York Stop Addictive Feeds Exploitation (SAFE) for Kids Act and implementing rulesNew York's law restricting algorithmic feeds for minors takes effect on 25 January 2027 under final rules published in July 2026, with certified age assurance and $5,000-per-violation penalties.

Knowledge Base Regulation Tracker Standards & Frameworks OSS Directory AI Incident Database Research Paper Library Case Studies Society Articles News Jobs About Charter Become a Member

Home/AI Incident Database/ChatGPT plugins made to exfiltrate conversation data through injected markdown image links

ChatGPT plugins made to exfiltrate conversation data through injected markdown image links

Johann Rehberger showed that a webpage summarised through a ChatGPT plugin could instruct the model to encode the user's chat history into an image URL that ChatGPT then fetched, leaking the data to the attacker's server.

Severity: Mediumchatgpt pluginsdata exfiltrationmarkdown injection

Sources

embracethered.comhttps://embracethered.com/blog/posts/2023/chatgpt-plugin-vulns-chat-with-code/

Contribute

Spot an error, or know a record we should add? Members can propose additions and submit corrections.

Submit incident Report a correction or request removal

ChatGPT plugins made to exfiltrate conversation data through injected markdown image links

Sources

Contribute

Related in AI Incident Database

"GrafanaGhost": prompt injection makes Grafana's AI assistant leak enterprise telemetry through image rendering

Copilot Studio agents exfiltrate data through retrieved documents (CVE-2026-21520) and the parallel "PipeLeak" flaw in Agentforce

Claude's Files API abused to exfiltrate user data through the code interpreter via prompt injection