monetr: Server-side request forgery in Lunch Flow link creation and refresh

Impact

A server-side request forgery (SSRF) vulnerability in monetr's Lunch Flow integration allowed any authenticated user on a self-hosted instance to cause the monetr server to issue HTTP GET requests to arbitrary URLs supplied by the caller, with the response body from non-200 upstream responses reflected back in the API error message.

The URL validator on POST /api/lunch_flow/link only checked the URL scheme and rejected query parameters; it did not filter loopback, RFC1918, link-local, or cloud-provider metadata addresses. The outbound HTTP client read the response body via an unbounded io.ReadAll, and the controller intentionally surfaced the resulting error (which contained the upstream body) as the JSON error field of the API response.

Who is affected: self-hosted monetr deployments running the default configuration. Out of the box, LunchFlow.Enabled=true, AllowSignUp=true, and billing is not enforced, so any user who can register on the instance can reach the vulnerable endpoint. Deployments running in a cloud environment where instance metadata is reachable from the pod (e.g. AWS EC2 without IMDSv2 enforced) expand the impact to include potential exposure of instance metadata through the reflected error body.

Who is NOT affected: the hosted my.monetr.app service, which runs with LunchFlow.Enabled=false. Self-hosted operators who had already disabled public sign-up (MONETR_ALLOW_SIGN_UP=false) substantially reduce their exposure since only operator-trusted users can reach the endpoint.

A secondary denial-of-service vector also existed: because the outbound response body was read with no size cap, an attacker-influenced upstream could return a multi-GB body that monetr would fully buffer into memory.

Patches

Fixed in monetr v1.12.5. Users should upgrade to this release or later.

The fix introduces a new config field LunchFlow.AllowedApiUrls (a list of permitted Lunch Flow API URLs) with a default of ["https://lunchflow.app/api/v1"]. URLs outside the allowlist are rejected both at link-creation time and at client-construction time, with a server-side warning log on rejection. Response body reads are capped at 10 MiB for both success and error paths. The UI renders the API URL field as a disabled pre-filled input when a single URL is allowed, or a dropdown when multiple are allowed, so operators who need to use a staging or self-hosted Lunch Flow API opt in explicitly via config.

Upgrade note for self-hosters with a custom Lunch Flow URL: if your existing LunchFlowLink records point at a URL other than , set your in your yaml config to include your custom URL before upgrading. Otherwise existing links will fail on next refresh or sync with a warning in the server log.