GHSA-g8wj-3cr3-6w7v is a medium-severity (CVSS 5.4) Cross-site Scripting (XSS) vulnerability in nuxt. O3 Security confirms whether GHSA-g8wj-3cr3-6w7v is actually reachable in your code before you act, and blocks exploitation at runtime until you patch.
Nuxt: `__nuxt_island` endpoint does not bind responses to request props, enabling shared-cache poisoning
Exploitation Status
No confirmed exploitation observed yet
- CISA’s own triage has not observed active exploitation or public proof-of-concept code for this CVE as of its last assessment.
Exploitation and automatability from CISA’s SSVC triage for GHSA-g8wj-3cr3-6w7v.
EPSS Exploitation Probability
EPSS (Exploit Prediction Scoring System) is a daily probability model maintained by FIRST.org. It estimates the likelihood a CVE will be exploited in production environments within the next 30 days, derived from real-world threat intelligence signals.
How urgent is this, really
GHSA-g8wj-3cr3-6w7v plotted by exploitation likelihood (EPSS) against impact (CVSS). The shaded corner — EPSS 50%+ and CVSS 7.0+ — is where this CVE doesn't sit, though severity or exploitability alone can still warrant action.
Where this sits among everything scored
Of 360,399 CVEs with a current EPSS score, this one falls in the < 10% band (highlighted). Real counts from FIRST.org, not a sample — log-scaled since the landscape is heavily right-skewed.
Real-World Exposure
How broadly this vulnerability is actually deployed: weekly install volume shows current usage, and reverse-dependency count shows how many other packages break if it stays unpatched.
nuxtnpm@nuxt/nitro-servernpmDescription
Summary
The /__nuxt_island/* endpoint accepts attacker-controlled props query/body parameters and renders any island component without verifying that the URL-resident hash (<Name>_<hashId>.json) was actually issued for those inputs by <NuxtIsland>. The hash is computed and embedded client-side but never validated server-side, so the same path can return materially different responses depending on the query.
Island components are documented as rendering independently of route context - page middleware does not apply to them, and they are intentionally cacheable as a function of their props. This advisory does not treat that contract as a vulnerability. It treats the absence of a binding between the URL the cache keys on and the response served at that URL as one.
Impact
In applications where a CDN or reverse-proxy in front of the app caches /__nuxt_island/* keyed by path only (ignoring query) - a documented misconfiguration class, see GHSA-jvhm-gjrh-3h93 - an attacker can prime the cache for a path with their own choice of props, and subsequent users requesting the same path receive the attacker's rendered HTML rather than the response intended for them. The cache entry persists until normal expiry.
Where the affected island has any prop flowing into an unsafe HTML sink in application code (v-html, innerHTML, a third-party renderer treating a prop as HTML), this becomes stored XSS in the embedding page's origin until the cache entry expires. HttpOnly cookies remain out of reach but anything else in the origin (other cookies, in-origin requests, DOM state) is reachable by the injected script.
Preconditions:
experimental.componentIslandsenabled (or the default'auto'with at least one server / island component in the app).- A shared intermediary cache (CDN, reverse-proxy, edge cache) keyed on path only.
- For the XSS pivot specifically: an application-authored island that puts a prop through an unsafe HTML sink.
Without the second precondition, the response shape is per-request and unaffected. Without the third, the worst case is content-swap / inert HTML injection rather than script execution.
Patches
Patched in [email protected] and [email protected] by #35077. The island handler now recomputes the expected hashId from (name, props, context) using the same ohash function <NuxtIsland> already uses to embed the hash in the URL, and rejects requests (HTTP 400) whose URL-resident hash does not match. The response is now a pure function of the request path: a path-keyed shared cache returns the correct response to every requester for that path, and an attacker cannot synthesise a path whose hash matches arbitrary props.
Workarounds
For users unable to upgrade immediately:
- Ensure any intermediary cache keys
/__nuxt_island/*on the full query string, not on the path alone. This is the recommended configuration regardless. - Audit application-authored islands for props flowing into
v-html/innerHTML/ similar HTML sinks; treat island props as untrusted user input.
Note on island authentication
[!IMPORTANT] It's important to remember that route middleware does not run when rendering island components, and islands cannot rely on routing-layer auth. Applications gating sensitive data behind page middleware should enforce that auth inside the island's own data layer (server-only routes,
useRequestEvent+ manual session checks, etc.) rather than relying on the embedding page's middleware - this was true before this advisory and remains true after it.
A separate advisory addresses *.server.vue pages registered as page_<routeName> islands, where the documented "middleware doesn't run for islands" contract collides with the page's own definePageMeta({ middleware }) declaration in a way that constitutes a genuine bug rather than documented behaviour.
Affected Packages
| Ecosystem | Package | Vulnerable range | Fix |
|---|---|---|---|
| 📦npm | nuxt | ≥ 3.1.0&&< 3.21.6 | 3.21.6 |
| 📦npm | nuxt | ≥ 4.0.0-alpha.1&&< 4.4.6 | 4.4.6 |
| 📦npm | @nuxt/nitro-server | ≥ 3.20.0&&< 3.21.6 | 3.21.6 |
| 📦npm | @nuxt/nitro-server | ≥ 4.2.0&&< 4.4.6 | 4.4.6 |
Detection & mitigation playbook
Open-source dependencyDetect
Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for nuxt. O3's reachability analysis confirms whether the vulnerable code path is actually invoked in your application, so you act on real exposure instead of every transitive match.
Fix
Update nuxt to 3.21.6 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms GHSA-g8wj-3cr3-6w7v is resolved across your whole dependency graph.
Workarounds
If you can't upgrade right away: gate or disable the affected feature, validate untrusted input at the boundary, and avoid passing attacker-controlled data into the vulnerable path. O3's runtime protection blocks exploitation in production as an interim safeguard until the upgrade lands.
How O3 protects you
O3 pinpoints whether GHSA-g8wj-3cr3-6w7v is reachable in your code and exactly where to fix it, then blocks exploitation in production at runtime until the patched version is deployed.
Tailored to GHSA-g8wj-3cr3-6w7v. Runtime protection reduces exposure until a permanent patch is applied and verified — it complements patching, it doesn't replace it.
Frequently Asked Questions
Is GHSA-g8wj-3cr3-6w7v in your dependencies?
O3 detects GHSA-g8wj-3cr3-6w7v across npm dependencies and uses function-level reachability to confirm whether the vulnerable code path is actually reachable — not just present. No false positives.