How to fix: Missing Origin-Agent-Cluster header
The server does not send Origin-Agent-Cluster: ?1, so the browser may place this page in a shared, document.domain-relaxable agent cluster instead of an origin-isolated one.
Why it matters
Without origin isolation, this page stays eligible for document.domain relaxation and shares its agent cluster (process, in Chromium) with same-site-but-different-origin pages, widening the blast radius of a Spectre-style side-channel and keeping the legacy document.domain same-origin downgrade available.
Origin-Agent-Cluster: ?1 asks the browser to give this exact origin its own agent cluster, separate from other origins on the same site. It also permanently disables document.domain relaxation for pages that send it, closing off a decades-old way for two subdomains to opt into treating themselves as same-origin.
How to fix it
- 1Add Origin-Agent-Cluster: ?1 to all responses.
- 2If any page relies on document.domain relaxation to communicate with a sibling subdomain, migrate that communication to postMessage before enabling this header, since it disables the relaxation for good on that origin.
Code examples
Next.js
// next.config.mjs
export default {
async headers() {
return [{ source: "/(.*)", headers: [{ key: "Origin-Agent-Cluster", value: "?1" }] }];
},
};Nginx
add_header Origin-Agent-Cluster "?1" always;References
Detected automatically by VulnRadar. The check id is origin-agent-cluster-missing, stable across every scan so you can filter or diff on it.
More headers checks
Questions
- How severe is Missing Origin-Agent-Cluster header?
- VulnRadar rates this info severity in the headers category. The severity is fixed per check, not scored per site, so the same finding is the same severity on every scan.
- Does VulnRadar check for this automatically?
- Yes. It runs as part of every scan under the check id origin-agent-cluster-missing, which stays the same between runs so you can filter, diff, or gate CI on it.
Missing a check?
The detection set is open. If you know a misconfiguration this does not catch, propose it: the issue template asks for the threat model, how to detect it, the false-positive risk, and a test fixture, which is the same thing a maintainer would have to work out anyway.