Skip to content
Noroxi
OSV ecosystem advisory

GHSA-7hgr-7h44-33w2: CamoFox MCP: Unauthenticated HTTP MCP browser-control surface

High—No exploit Fix available

Noroxi assessment

Priority: Monitor (action score 0/100)

  • The sources we track (CISA KEV, Exploit-DB, Metasploit, Nuclei, GitHub) show no public exploit code or exploitation in the wild.
  • Anyone who can reach it over the internet can trigger it. No account or password is required. No user action is required. No special conditions are required; it is repeatable.
  • Fixed version: camofox-mcp (npm) 1.13.2.

Low priority for now. Upgrade to camofox-mcp (npm) 1.13.2 or later.

Generated automatically from the record's data: CVSS (NVD and CNA), CISA KEV and SSVC, FIRST EPSS, exploit sources, fixed versions. How we calculate it

Primary sources: OSV

# Unauthenticated HTTP MCP browser-control surface in `camofox-mcp` ## Summary `camofox-mcp` exposed a Streamable HTTP MCP endpoint at `/mcp` with rate limiting but no inbound MCP-layer authentication. When HTTP mode was enabled, any client that could reach `/mcp` could list and invoke browser-control tools. If `CAMOFOX_API_KEY` was configured, the server then forwarded that server-side key to the underlying `camofox-browser` backend. That means an unauthenticated MCP caller could exercise the server's browser authority without knowing the backend browser API key. Reviewed vulnerable commit: `10e3ac08cb50d830eb4ee00a789229f02f28a1a4` Fixed commit observed on main: `599f56ee40f8062aeca541c251ed1d39fb437f50` Fixed release observed: `v1.13.2` Suggested severity: High, with the caveat that default loopback-only deployments reduce practical exposure. ## Root cause In the reviewed commit, `src/http.ts` creates the Express MCP app and applies only a rate limiter to `/mcp`: ```ts const app = createMcpExpressApp({ host: config.httpHost }); const limiter = rateLimit({ windowMs: 60_000, limit: config.httpRateLimit, standardHeaders: true, legacyHeaders: false }); app.use("/mcp", limiter); ``` The `POST /mcp` handler then creates a server and `StreamableHTTPServerTransport` and passes the request body into the MCP transport without checking `Authorization`, an inbound API key, allowed hosts, or public-bind safety: ```ts app.post("/mcp", async (req: any, res: any) => { try { const { server } = createServer(config); const transport = new StreamableHTTPServerTransport({ sessionIdGenerator: undefined }); await server.connect(transport); await transport.handleRequest(req, res, req.body); ``` `src/config.ts` made HTTP mode configurable and defaulted the HTTP host to loopback, but it did not require an inbound HTTP client secret: ```ts transport: cli.transport ?? envTransport ?? "stdio", httpPort: cli.httpPort ?? (Number.isNaN(httpPortFromEnv) ? 3000 : httpPortFromEnv), httpHost: cli.httpHost ?? env.CAMOFOX_HTTP_HOST ?? "127.0.0.1", ``` Separately, `src/client.ts` forwarded `CAMOFOX_API_KEY` server-side to the browser backend: ```ts if (this.apiKey) { headers.set("x-api-key", this.apiKey); headers.set("authorization", `Bearer ${this.apiKey}`); } ``` So `CAMOFOX_API_KEY` protected the MCP server's outbound requests to the backend browser service, but did not authenticate inbound HTTP MCP clients. ## Auth boundary The vulnerable boundary was the HTTP MCP endpoint. The client did not need to provide `Authorization` or any `CAMOFOX_API_KEY` value to call MCP tools. The default bind was `127.0.0.1`, which lowers severity for default local-only deployments. The risky cases are documented HTTP/remote-client deployments, Docker/port-forwarded deployments, or any environment where a browser page, local network client, reverse proxy, or another user can reach the `/mcp` endpoint. ## Proof of concept I used a fake `camofox-browser` backend so no real browser was launched and no external navigation occurred. The harness starts the reviewed `dist/http.js` server with `CAMOFOX_API_KEY=server-side-secret`, connects an MCP SDK client to `/mcp` with no auth headers, lists tools, then calls `create_tab` and `navigate`. Observed output: ```json { "authUsedByClient": false, "listedToolCount": 46, "backendRequests": [ { "method": "POST", "url": "/tabs", "headers": { "authorization": "Bearer server-side-secret", "x-api-key": "server-side-secret" } }, { "method": "POST", "url": "/tabs/fake-tab-1/navigate", "headers": { "authorization": "Bearer server-side-secret", "x-api-key": "server-side-secret" } } ], "observedUnauthenticatedBrowserControl": true, "serverSideSecretForwardedToBackend": true } ``` This demonstrates both parts of the issue: 1. The MCP client used no inbound authentication. 2. The server still used its configured backend browser secret when forwarding the tool calls. ## Impact An unauthenticated client that can reach the HTTP MCP endpoint can exercise browser-control tools as the MCP server. Depending on the user's active browser profiles and configured backend, that can allow page navigation, tab creation, interaction with authenticated browser contexts, screenshot/content observation, and other browser-automation actions exposed by the MCP tool surface. The impact is strongest when HTTP mode is intentionally exposed for remote MCP clients or through Docker/reverse-proxy deployment and the operator assumes `CAMOFOX_API_KEY` protects the whole control plane. ## Fix notes The public issue indicates this has been fixed in `599f56e` and released as `v1.13.2` by adding dedicated inbound `CAMOFOX_HTTP_API_KEY` Bearer auth, public-bind startup validation, auth before `/mcp` JSON parsing, loopback Host-header protection, and optional allowed-hosts handling. Those are the right mitigation directions.

Published
May 19, 2026
Updated
May 19, 2026
EPSS
—

Report tools

JSON

Action score

0

Monitor

Low priority for now.

CVSS
0 / 40 · —
CISA KEV
0 / 30 · Not listed
EPSS
0 / 30 · —

Get told when this record changes

  • Not in KEV: no record of exploitation in the wild
  • No public exploit code on record
  • A fix has been released

If any of these changes, we'll tell you in the daily email digest.

Follow GHSA-7hgr-7h44-33w2

When you'll hear from us

  • when it enters CISA KEV
  • when exploit code appears or is weaponized
  • when the record is updated (including a fixed version)
  • when EPSS rises 20 points in a week

One daily email digest (06:00 UTC); connect a webhook for KEV listings and weaponized exploits every 10 minutes on paid plans, on the hour on the free plan.

Your free account also gets you

  • CVE alerts: add the products and versions you run, get the new CVEs that affect them
  • Uptime monitoring: 1 site, instant email when it goes down
  • Add your domain: subdomains and takeover risk scanned every week (1 domain)
  • 10 scans an hour on the security tools, at most 30 a day (3 a day without an account), an API key and MCP
  • All your scans in one panel: findings, what changed in the last scan, a report per domain or per client (PDF, share link)
All free tools
Is your version affected? Check it

Noroxi analysis

What does it mean in your environment?

The assessment above is generated automatically from the record's data; it doesn't know your environment. Our team verifies on your systems whether this vulnerability is actually exploitable and which mitigation is enough.

We use this product, ask for help

Affected systems

Is your version affected?

Pick the product and type the version you run. It's compared with this record's version ranges in your browser; the version isn't sent anywhere.

VendorProduct
npmcamofox-mcp

Package-level exposure

OSV and GitHub Advisory data: ecosystem, package and range. SBOM matching uses this table.

EcosystemPackageAffected rangeFix
npmcamofox-mcpbefore 1.13.21.13.2

Remediation

Which version to upgrade to

Fix versions compiled from the vendor, package registries and Microsoft. Verify the vendor's note before upgrading.

Product / packageFixed versionSource
npm:camofox-mcp1.13.2Package registry (OSV)

Exploit status

No known public exploit

No public exploit has been observed yet. That doesn't mean you're safe, only that the bar is a little higher.

Research context

For pentesters and researchers: attack profile, score disagreement, timeline, patch commits, credits, variant and chain candidates, bug bounty scope. All derived from existing data; no exploit code.

Timeline

From publication to today: proof of concept, Metasploit module, CISA KEV and fix record. Dates are as reported by the sources.

No dated events beyond publication.

Patch and commit links

Commit, PR and diff links among the references. A starting point for patch-diffing and variant hunting; fixes, not exploits.

Not on record for this CVE: EPSS history · researcher credit · variant or chain candidate · public bounty program.

Technical details

Attack conditions

  • Anyone who can reach it over the internet can trigger it.
  • No account or password is required.
  • No user action is required.
  • No special conditions are required; it is repeatable.

If successful

Confidentiality
Integrity
Availability
Attack vector
Network
Attack complexity
Low
Privileges required
None
User interaction
None

CVSS vector

CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:L/VI:N/VA:N/SC:H/SI:H/SA:L

osv

Root cause

A function added for development or support ships to production without authentication. It often relies on the assumption that it is “only reachable from the internal network.”

A representative example of this vulnerability class. Not the vendor's source code; it shows the faulty pattern and its fix.

Vulnerable

ts
// Added for the support teamapp.post("/maintenance/reset", resetDevice);

Fixed

ts
if (config.maintenanceEnabled) {  app.post(    "/maintenance/reset",    requireSession, requireRole("admin"), resetDevice  );}

Attack context

MITRE CAPEC attack patterns and ATT&CK techniques for this weakness class (CWE). A starting point for detection rules and threat hunting.

References

Show all references (2 more)

Vendor advisories and official records. Exploit/PoC links are deliberately left out.

All records