The Pulse
Antigen Puts tCell on the Attack—and on the Hook for Fixes
Antigen’s tCell agent is designed to find and exploit vulnerabilities, propose patches for engineers to review, and retest after deployment. The company describes managed and self-hosted deployment options for running the system.
AI.info Team · Written with AI ·
Antigen’s tCell agent is designed to attack a company’s systems, draft fixes for vulnerabilities it finds and then try the exploits again after engineers deploy those fixes. Founderland reported the product launch on October 9, 2026; Antigen’s website describes a workflow that links offensive testing with human-reviewed remediation and retesting.
“We build an AI agent that hacks into you, finds security vulnerabilities and fixes them.”
Saad Jamal, Antigen co-founder, speaking in a video
From an exploit to a proposed patch
Antigen says tCell maps an organization’s exposed systems, signs into accounts and links weaknesses into attack paths. The company says each finding includes a reproducible exploit, which gives security teams evidence of how a flaw might be used rather than only a scanner alert.
After identifying a vulnerability, the platform writes a proposed fix and opens a pull request. Antigen says the patch goes to engineers for review and approval; the system does not merge it or deploy it on its own. The company lists GitHub and Linear integrations for this workflow.
Retesting is part of the product claim
Antigen’s advertised sequence continues after a fix reaches production: tCell attempts the original exploit again. If that test surfaces another issue, the company says the platform can reopen the finding with logs and draft another patch informed by the result.
That loop distinguishes the product’s pitch from a conventional point-in-time test, which delivers findings that engineering teams must address and verify separately. Antigen says its reports are reviewed by a security researcher before they reach customers, adding a human check alongside the automated testing.
Testing scope and deployment choices
The company presents two deployment options. Its managed service runs on Antigen-managed infrastructure, while its Dedicated option places self-hosted workers inside a customer’s environment, including private networks on AWS, Azure, Google Cloud or Kubernetes. Antigen says Dedicated deployments keep findings, logs and artifacts on the customer’s systems.
Those controls matter because an agent that probes live systems can cause harm if it tests the wrong target or crosses an approved boundary. Antigen’s public product pages describe policy controls and deployment choices, but provide limited detail about the testing safeguards themselves. Teams considering the product would need to establish which targets and actions are authorized before running it.
Product claims still need outside evidence
Antigen describes tCell as a way to keep testing aligned with frequent software changes, and its site offers a live demonstration against a customer’s infrastructure.
That leaves an important distinction for buyers: the site explains the intended workflow, but does not establish how often tCell finds exploitable flaws, how reliably its patches resolve them, or how much human review each case requires. For now, the concrete promise is a proposed patch awaiting engineer approval, followed by another attempt to reproduce the original exploit.