Best possible solution:

Land the real-transport QA evidence path once maintainers accept the automation cost and current-head required checks remain green.

Do we have a high-confidence way to reproduce the issue?

Not applicable: this PR is a QA coverage/evidence change, not a reported product bug. The PR body and current-head proof check provide real scenario producer evidence instead of a bug reproduction path.

Is this the best way to solve the issue?

Yes: using the existing QA Gateway child fixture, script evidence writer, and MCP/Gateway transport surfaces is the right boundary for proving these scenarios without changing product runtime behavior.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against fd0355d5b1a1.

Label changes:

• add merge-risk: 🚨 automation: The diff changes QA scenario execution to real child-process, WebSocket, and MCP stdio script automation that can affect CI/proof reliability after merge.
• add proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes terminal-style real scenario producer results and the current-head Real behavior proof check completed successfully.
• add rating: 🦞 diamond lobster: Overall readiness is 🦞 diamond lobster; proof is 🦞 diamond lobster and patch quality is 🦞 diamond lobster.
• add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (terminal): The PR body includes terminal-style real scenario producer results and the current-head Real behavior proof check completed successfully.

Label justifications:

• P3: This is a low-risk maintainer QA evidence improvement with no intended product behavior change.
• merge-risk: 🚨 automation: The diff changes QA scenario execution to real child-process, WebSocket, and MCP stdio script automation that can affect CI/proof reliability after merge.
• rating: 🦞 diamond lobster: Overall readiness is 🦞 diamond lobster; proof is 🦞 diamond lobster and patch quality is 🦞 diamond lobster.
• status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (terminal): The PR body includes terminal-style real scenario producer results and the current-head Real behavior proof check completed successfully.
• proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes terminal-style real scenario producer results and the current-head Real behavior proof check completed successfully.

PR surface:

Source +85, Tests +723, Other +17. Total +825 across 8 files.

What I checked:

• Live PR state: Live GitHub metadata shows the PR is open, authored by a repository member, labeled maintainer, mergeable, and currently at head 02d017f. (02d017f5cd8a)
• Real transport producer: The added producer starts a real QA Gateway child, runs the existing Gateway smoke client against its WebSocket URL, and checks Gateway health over RPC. (test/e2e/qa-lab/runtime/gateway-mcp-real-transports.ts:416, 02d017f5cd8a)
• Startup retry proof path: The producer starts an MCP stdio client through a WebSocket recording proxy before Gateway readiness and asserts startup-unavailable retries, connect protocol range, and negotiated protocol. (test/e2e/qa-lab/runtime/gateway-mcp-real-transports.ts:453, 02d017f5cd8a)
• Plugin-tools proof path: The producer starts the real plugin-tools stdio server, lists tools with the MCP SDK client, and calls the fixture plugin's memory_search tool. (test/e2e/qa-lab/runtime/gateway-mcp-real-transports.ts:570, 02d017f5cd8a)
• Gateway child hook placement: The PR-head Gateway child fixture waits for /healthz listening, invokes onListening with URL/token/config context, then waits for full readiness before creating the QA RPC client. (extensions/qa-lab/src/gateway-child.ts:904, 02d017f5cd8a)
• Current MCP retry contract: Current main treats retryable GatewayClientRequestError values and connect timeout messages as recoverable during initial MCP Gateway connect. (src/mcp/channel-bridge.ts:651, fd0355d5b1a1)

Likely related people:

• steipete: Peter Steinberger has the largest path-history footprint across the QA/MCP files sampled and introduced the repo-backed QA suite runner in commit 4780788. (role: feature-history owner; confidence: high; commits: 4780788bbbf9, fcee2683736a, 610407730d42; files: extensions/qa-lab/src/gateway-child.ts, qa/scenarios/runtime/gateway-smoke.yaml, src/mcp/plugin-tools-serve.ts)
• RomneyDa: Dallin Romney authored this PR and has recent merged/open QA transport work, including prior Gateway/MCP scenario coverage and QA smoke reliability work found in live PR search. (role: recent QA transport contributor; confidence: medium; commits: 02d017f5cd8a; files: test/e2e/qa-lab/runtime/gateway-mcp-real-transports.ts, extensions/qa-lab/src/gateway-child.ts, qa/scenarios/runtime/mcp-gateway-connect-startup-retry.yaml)
• vincentkoc: Vincent Koc appears repeatedly in MCP/plugin-tools path history and current MCP plugin-tools PR search results, including commits around runtime seams and release QA stabilization. (role: MCP adjacent owner; confidence: medium; commits: 0f7d9c957093, 74e7b8d47b18, e085fa1a3ffd; files: src/mcp/channel-bridge.ts, src/mcp/plugin-tools-serve.ts, qa/scenarios/plugins/mcp-plugin-tools-call.yaml)
• gumadeiras: Gustavo Madeira Santana has multiple QA Lab staging and scenario organization commits in the touched QA helper area. (role: QA staging contributor; confidence: medium; commits: 82fe6f50efc4, bb7e9823a886, d5933af80bf1; files: extensions/qa-lab/src/bundled-plugin-staging.ts, extensions/qa-lab/src/gateway-child.ts, qa/scenarios/runtime/gateway-smoke.yaml)

• 🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
• 🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
• 🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
• 🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
• 🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
• 🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
• 🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

• ClawSweeper keeps one durable marker-backed review comment per issue or PR.
• Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
• A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
• PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
• Maintainers can also comment @clawsweeper review to request a fresh review only.
• Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
• Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
• Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.