Skip to content

Validate upstream JSON-RPC responses in transparent proxy#5288

Open
bishnubista wants to merge 1 commit into
stacklok:mainfrom
bishnubista:fix/validate-upstream-jsonrpc-responses
Open

Validate upstream JSON-RPC responses in transparent proxy#5288
bishnubista wants to merge 1 commit into
stacklok:mainfrom
bishnubista:fix/validate-upstream-jsonrpc-responses

Conversation

@bishnubista
Copy link
Copy Markdown

@bishnubista bishnubista commented May 14, 2026

Summary

The transparent proxy (thv run) currently forwards upstream MCP server responses to clients with HTTP 200 even when the JSON-RPC frame is structurally invalid — missing jsonrpc:"2.0", invalid id type, non-object body, etc. The proxy sits between potentially untrusted upstream MCP servers and trusted clients, so a compromised or misconfigured upstream can push malformed frames straight through, where they may crash strict JSON-RPC parsers, be misinterpreted as legitimate responses, or trigger undefined behaviour in client-side parsers.

This PR validates upstream JSON-RPC frames at the proxy boundary and rewrites invalid responses into a structured JSON-RPC error so the proxy stops being a silent amplifier for malformed (or adversarial) upstream servers.

  • Validation runs in NoOpResponseProcessor.ProcessResponse (streamable-http / default transport). SSE is unaffected.
  • Structural rules: single JSON value (trailing JSON rejected); object or non-empty array of objects; per object jsonrpc=="2.0", id ∈ {string, number, null}, exactly one of result/error; error.code must be an integer, error.message must be a string.
  • Gating: POST + 200, application/json(-rpc) media type, non-identity Content-Encoding passed through untouched, request must carry an MCP streamable-HTTP signal (MCP-Protocol-Version or Mcp-Session-Id) so non-MCP JSON traffic flowing through the catch-all is not rewritten.
  • Invalid frame → HTTP 502 with body {"jsonrpc":"2.0","id":null,"error":{"code":-32000,"message":"Invalid upstream JSON-RPC response","data":"<detail>"}}. -32000 is the JSON-RPC implementation-defined server-error code; -32603 is reserved for internal JSON-RPC errors. Response headers are replaced wholesale so upstream Mcp-Session-Id, Set-Cookie, ETag, Cache-Control, etc. are not smuggled into the proxy-generated error.
  • Body reads bounded by io.LimitReader at maxJSONRPCResponseBytes = 100 << 20 (100 MiB), matching streamable-HTTP precedent in pkg/vmcp/client and pkg/vmcp/session/internal/backend. Oversized responses are rejected with the same 502 shape so the proxy cannot be amplified into a memory DoS.

Closes #5247

Type of change

  • Bug fix
  • New feature
  • Refactoring (no behavior change)
  • Dependency update
  • Documentation
  • Other (describe):

Test plan

  • Unit tests (task test)
  • E2E tests (task test-e2e)
  • Linting (task lint-fix)
  • Manual testing (describe below)

Local verification:

  • go test ./pkg/transport/proxy/transparent/... -count=1 — pass.
  • golangci-lint run --timeout=5m ./pkg/transport/proxy/transparent/... — 0 issues.

New test cases in pkg/transport/proxy/transparent/response_processor_test.go:

  • Valid pass-through: result, error, batch, result: null, application/json; charset=utf-8.
  • Invalid → 502: missing jsonrpc, invalid id type, non-object body, result+error both present, trailing JSON, fractional error.code.
  • Non-validated pass-through: non-POST, non-200, non-JSON, text/event-stream, application/jsonsomethingelse.
  • Compressed responses (Content-Encoding: gzip) passed through unchanged for both valid and malformed payloads; explicit Content-Encoding: identity still validates.
  • MCP-signal gate: no MCP request headers → pass through even with malformed body; MCP-Protocol-Version or Mcp-Session-Id → validate.
  • Rewritten 502 strips Mcp-Session-Id, Set-Cookie, Etag, Cache-Control from the response.
  • Oversized response (> maxJSONRPCResponseBytes) is rejected with a size-limit error.

API Compatibility

  • This PR does not break the v1beta1 API, OR the api-break-allowed label is applied and the migration guidance is described above.

No CRD or operator API surface is touched.

Does this introduce a user-facing change?

Yes. Clients of the transparent proxy that previously received malformed upstream responses verbatim with HTTP 200 will now receive HTTP 502 with a synthetic JSON-RPC error body. Conformant upstream MCP servers are unaffected — only structurally invalid responses are rewritten.

Special notes for reviewers

Two scope decisions called out explicitly so reviewers can push back if either is wrong for the project:

  1. Validation runs in ModifyResponse, after tracingTransport.RoundTrip may have observed an upstream Mcp-Session-Id and registered proxy-side session state. This PR strips upstream session headers from the proxy-generated 502 so clients never receive a session id derived from a malformed initialize response, but it does not roll back server-side proxy session state created before validation. Moving validation earlier (or adding a rollback path for invalid initialize responses) touches tracingTransport and felt outside the scope of "structurally validate the upstream frame." Happy to open a follow-up — or fold it into this PR if maintainers prefer.

  2. Backward-compat clients that POST without MCP-Protocol-Version on the very first initialize will not trigger validation — same as today's behaviour, so no regression. If broader coverage is preferred, tracingTransport.RoundTrip already buffers and parses the request body to detect initialize; a small context-propagated marker would let ProcessResponse validate those frames too. Easy follow-up.

Open to feedback on:

  • The 100 MiB body cap (matched pkg/vmcp precedent; can tighten to 10 MiB if maintainers prefer an explicit security default).
  • Whether the MCP-signal narrowing should be widened to also accept the buffered-request-body sniff above.

The transparent proxy forwarded malformed upstream MCP frames to clients with
HTTP 200 even when the response violated JSON-RPC 2.0 structure. This adds a
boundary check in NoOpResponseProcessor that rejects structurally invalid
upstream frames and returns a synthetic 502 carrying a JSON-RPC error to the
client, so the proxy stops being a silent amplifier for malformed (or
adversarial) upstream servers.

Validation runs only for streamable-http POST/200 responses that carry an MCP
request signal (MCP-Protocol-Version or Mcp-Session-Id) and an application/json
content type, with non-identity Content-Encoding traffic passed through
untouched. Body reads are bounded to 100 MiB to match existing streamable-HTTP
limits in pkg/vmcp. Rewritten error responses replace headers wholesale so
upstream session/cookie/cache metadata is not smuggled into the proxy-generated
error. SSE traffic is unaffected.

Closes stacklok#5247

Signed-off-by: bishnubista <bista.developer@gmail.com>
@bishnubista bishnubista force-pushed the fix/validate-upstream-jsonrpc-responses branch from 4fcdb4b to a7badd9 Compare May 14, 2026 23:58
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

security: proxy forwards malformed upstream JSON-RPC frames to clients without validation

1 participant