Quick answers
- Is CVE-2026-120233 exploited in the wild?
- No. Junglewise Threat Intelligence has no report of CVE-2026-120233 being exploited in the wild.
- How severe is CVE-2026-120233?
- Medium severity, with a CVSS score of 5.3, published 1 November 2026.
- Which products and versions are affected?
- Ag-Ui-Protocol Ag-Ui up to 0.0.58.
- Is there a fix for CVE-2026-120233?
- No fixed version is recorded yet. Set redirect: "manual" or redirect: "error" in the HttpAgent's requestInit to disable automatic redirect following and treat redirect responses as errors instead of forwarding the credential-bearing request body.</anigation> <parameter name="timeline">[{"kind":"disclosed","at":"2026-11-01"},{"kind":"other","note":"Public exploit released; fix awaiting pull request acceptance"}] If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.
- How is CVE-2026-120233 attacked?
- It is attacked over the network, in a software library.
- Am I affected by CVE-2026-120233?
- Check whether you run Ag-Ui-Protocol Ag-Ui at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.
How to fix
Set redirect: "manual" or redirect: "error" in the HttpAgent's requestInit to disable automatic redirect following and treat redirect responses as errors instead of forwarding the credential-bearing request body.</anigation> <parameter name="timeline">[{"kind":"disclosed","at":"2026-11-01"},{"kind":"other","note":"Public exploit released; fix awaiting pull request acceptance"}]
Technical details
The vulnerability exists in the HttpAgent's requestInit configuration, which omits the redirect option, causing the Fetch API to default to "follow" and automatically re-issue requests to attacker-chosen URLs. When a 307/308 redirect is returned, the method and body are preserved, and developer-supplied headers (X-Tenant, API keys) are forwarded across origins. The attack requires a malicious or compromised agent endpoint but is remotely exploitable and can exfiltrate credentials and run-request envelopes (threadId, runId, conversation data) to arbitrary origins or perform client-side request forgery against internal endpoints.
Affected products
- ag-ui-protocol ag-ui up to 0.0.58