Data exfiltration
TrialTechniques
The unauthorized extraction of sensitive data from a system.
Why it's here
Placed in Trial: 6 article(s) of evidence from 3 source(s), led by security coverage, with 3 in the last 30 days. Confidence 64%.
Evidence (6)
- 6Hacker News·6/12/2026regulationDigital sovereignty gains urgency over email surveillance fears
The article argues that concerns about foreign access to Dutch email data are making digital sovereignty a more immediate policy priority. It frames the issue as a warning for governments and organizations to reduce dependence on foreign-controlled services and infrastructure.
- 8The New Stack·6/11/2026securityChainguard flags greyware in open-source packages
Chainguard says its new source code scanner is detecting not only malware but also “greyware” in public registries, meaning packages that behave as advertised while also including harmful capabilities. The company says it has scanned over 100,000 packages per day and blocked more than 52,000 packages identified as malware or greyware, with examples on npm involving credential theft, token exfiltration, and persistent backdoors.
- 7The New Stack·6/11/2026product_launchlakeFS targets safer agent writes to production data
lakeFS announced a new service for agentic AI aimed at providing governed, reproducible access to enterprise production data. The company argues that manual data stewardship cannot keep up when autonomous agents make parallel writes at machine speed, increasing the risk of irreversible corruption without isolation and rollback controls.
- 7OpenAI Blog·3/25/2026securityOpenAI launches Safety Bug Bounty program
OpenAI has introduced a Safety Bug Bounty program to help identify AI abuse and safety risks in its systems. The program will focus on issues such as agentic vulnerabilities, prompt injection, and data exfiltration.
- 7OpenAI Blog·2/13/2026securityChatGPT adds Lockdown Mode and Elevated Risk labels
OpenAI is introducing Lockdown Mode and Elevated Risk labels in ChatGPT to help organizations better defend against prompt injection and AI-driven data exfiltration. The new controls are designed to make security-sensitive usage of ChatGPT easier to manage in enterprise settings.
- 5OpenAI Blog·1/28/2026securityOpenAI details link-click safeguards for AI agents
OpenAI explains how it protects user data when AI agents open links, with safeguards designed to reduce the risk of URL-based data exfiltration and prompt injection. The guidance focuses on built-in protections that help keep agent-driven browsing safer.