OpenAI Daybreak adds agents for testing and vulnerability fixes
Original title: OpenAI’s Daybreak brings AI agents to security testing and vulnerability fixes
Early-access Red Team confirms flaws and retests patches, while Defense Factory drafts fixes engineers review before merge.
OpenAI introduced Daybreak, pairing coordinated agents for web-application security testing with agents that investigate findings and prepare patches. Red Team, in an early-access program, reviews source and context, checks scanner, bug-bounty, and disclosure reports, and retests whether a patch blocks the issue. Defense Factory traces findings through code and proposes fixes; engineers review them before merging. Teams install a ChatGPT preview plugin, connect GitHub, and can scan cloud or local checkouts.
- Assessments use isolated environments and Guardian review to limit reachable endpoints.
- Cloud scans run in Codex cloud and require a personal GitHub connection.
- Scans can cover a default branch, selected paths, new commits, or pull requests.
- An optional Deep scan mode makes multiple passes and uses more tokens.
- Vendors
- OpenAI
From the original article
1 min read at helpnetsecurity.com
OpenAI is bringing AI agents into more of the software security process with Daybreak, pairing web application penetration testing with tools that investigate vulnerabilities and develop fixes. The two offerings, Red Team and Defense Factory, connect the work of finding security weaknesses with patching and retesting them.
By Anamarija Pogorelec · helpnetsecurity.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.