Author	SHA1	Message	Date
Cliff Hall	53b619db09	Merge pull request #1958 from modelcontextprotocol/v2/fix/unhandled-rejections-progress-tests ... fix: settle the deliberately in-flight tool calls in the progress tests	2026-08-11 11:17:55 -04:00
cliffhall	8049b4f7c8	test: cover the structured-output cap and the list_items showcase end to end ... Two gaps from the Copilot review on #1975: - The `Large` story asserted nothing, so dropping the section's `mah` (or applying it to the wrong element) would still pass CI. Its play function now measures the real ScrollArea viewport: the visible height stays at the cap and the overflow is scrollable rather than clipped. - Nothing exercised the `list_items` fixture, so a preset-registration typo would leave the documented showcase broken with a green suite. A new integration test loads `structured-output-http.json` through loadConfig/resolveConfig — covering the file, the registry entry, and the fixture — connects, calls the tool, and asserts both the summary block and the nested structuredContent. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SW1p8E2uiyyLx4RKwrwSrt	2026-08-11 10:28:01 -04:00
cliffhall	347173fab7	fix: absorb only the teardown's own close, not any ConnectionClosed ... Review feedback (round 3). The predicate absorbed *any* CONNECTION_CLOSED, while the doc comment claimed it absorbed only the teardown's — so a transport dropping on its own after the progress notifications but before `disconnect()` would have been swallowed. These tests assert only on the notifications, so that regression would have passed green. Gate the absorption on a teardown flag owned by the helper, which now returns an `InFlightCall` exposing `disconnectAndSettle()`. Keeping the flag inside means it can't be raised too early (which would silently reopen the hole) and the settle can't be forgotten; call sites also collapse from two statements to one. Verified: a pre-teardown ConnectionClosed now fails the test and stays observed (no run-level error), while a teardown-time one is still absorbed. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HDdo1rNRVQnVRRdsqSrvG3	2026-08-10 18:19:33 -04:00
cliffhall	c60563b7d6	fix: observe the derived promise so an unexpected rejection can't go unhandled ... Review feedback (round 2). Narrowing the handler in the previous commit fixed the silent-pass but reintroduced #1947's own failure class on the unexpected path: `then` returns a *derived* promise, and the re-throw rejects that one rather than `call`. The caller doesn't await it until after `disconnect()`, so a rejection arriving while the test is still waiting on progress notifications sits unobserved for seconds and Node reports it as an unhandled rejection. Attach an observer to the derived promise the moment it exists. Nothing is swallowed — it stays rejected, so the caller's `await` still fails the test. Verified both directions with a probe rejecting with a non-ConnectionClosed error and awaiting after a 1.5s gap: without the observer the run reports "Unhandled Rejection" + "Errors 1 error"; with it, the test fails and the run stays clean. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HDdo1rNRVQnVRRdsqSrvG3	2026-08-10 18:09:52 -04:00
cliffhall	0b9aee11a3	fix: only absorb the teardown's own rejection in settleInFlight ... Review feedback: the blanket rejection handler also converted a genuine `callTool` failure into a pass. That matters here because these tests assert on the progress notifications, which arrive before the result — so a call that emitted all three progress events and then failed would have satisfied every assertion and been swallowed. Narrow the handler to absorb only `SdkError` / `CONNECTION_CLOSED` and re-throw anything else. The handler is still attached at call time, so an unexpected rejection stays *handled* and surfaces as an ordinary test failure at the `await inFlight` after `disconnect()` rather than as another unhandled rejection. Fulfillment is still accepted: whether the call beats the teardown is a race, so asserting either outcome would reintroduce the original bug as a flake. Verified the narrowed predicate is actually exercised rather than dead code — both call sites reject with SdkError/CONNECTION_CLOSED on every run. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HDdo1rNRVQnVRRdsqSrvG3	2026-08-10 18:01:27 -04:00
cliffhall	1881e5c73f	fix: settle the deliberately in-flight tool calls in the progress tests ... Two progress tests fire `client.callTool()` without holding the promise so they can assert on the notifications the call streams while it is still in flight. `disconnect()` then closes the SDK client, which rejects every pending request with "Connection closed" — with nothing holding the promise that lands as an unhandled rejection, which vitest counts as a run error and fails `npm run ci` at `coverage` even though all tests pass. Add a `settleInFlight()` helper that attaches the handler at call time (not after the assertions, which would leave a window the rejection can escape through) and returns a promise the test awaits after `disconnect()` so teardown stays ordered. Neither outcome is asserted: the call may legitimately complete before the teardown or reject with it. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HDdo1rNRVQnVRRdsqSrvG3	2026-08-10 17:02:05 -04:00
Cliff Hall	c2af547cc8	fix(core): register peer-request handlers before connect (#1797) (#1798) ... * fix(core): register peer-request handlers before connect (#1797) The roots/sampling/elicitation capabilities are advertised on the SDK `Client` at construction, so from the moment `connect()` sends `notifications/initialized` the server may issue any of those requests. The handlers, however, were registered after the handshake resolved and after `fetchServerInfo()` — leaving a window in which the SDK answered `-32601 Method not found`. `server-filesystem` asks for `roots/list` the instant it is initialized (that is how it learns its allowed directories) and lost the race, so it silently fell back to its CLI-argument directories instead of the roots configured in the Inspector. `server-everything` asks later and did not. Extract the four peer-request handler blocks into `registerPeerRequestHandlers()` and call it before `client.connect()`. None of them depend on server capabilities — only on constructor-set state. The notification handlers that do gate on `this.capabilities` stay in `connect()`, after `initialize`. Regression test drives the race deterministically with a fake transport that delivers `roots/list` synchronously from inside the `send()` of `notifications/initialized`; it fails on the pre-fix ordering. Closes #1797 Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-1 fixes — pre-connect roots/list_changed, wider timing test (#1797) Finding 1: `notifications/roots/list_changed` was the one handler in the post-handshake block that gates on nothing but constructor state, so it sat in the same race — a server emitting it right after `initialized` had it silently dropped (no wire error; the symptom is a missed UI refresh). Moved to a new `registerPeerNotificationHandlers()` called next to `registerPeerRequestHandlers()` before `connect()`. Finding 2: the timing test asserted one of the four moved handler blocks. It now injects `tasks/list` and the `roots/list_changed` notification in the same burst as `roots/list`, so moving any of the block back after `connect()` fails the test. Sampling/elicitation stay out — they park a pending request awaiting user input and have no reply to assert on. Dropped the unused `sessionId`/`setProtocolVersion` transport fields. Verified the added assertions are load-bearing: the notification assertion fails against the previous commit, the `roots/list` one against `origin/v2/main`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-2 fixes — correct the roots/list_changed premise (#1797) The round-1 doc comment claimed a server may emit `notifications/roots/list_changed` at initialize. It can't, normally: `RootsListChangedNotificationSchema` is a member of `ClientNotificationSchema` and absent from `ServerNotificationSchema` — the client sends it (setRoots does, at :4251). The inbound handler is defensive coverage for a non-conformant server, and dispatches `rootsChange` with our own already-known roots. Reworded the method doc and the test's header so neither states the false premise; the move itself stands (it gates on no server capability, so it belongs with the request handlers). Also from review: - Dropped the vacuous `notificationRejected` flag and its inverted comment. The SDK does not throw for an unregistered notification, so the try/catch could never observe one — confirmed by bisect, where the pre-fix failure was always `rootsChanges`, never the flag. Removing the catch also stops it swallowing a genuine throw from the injected requests. - Narrowed `replies` to responses (`"result" in message || "error" in message`) so it can't also collect client-originated requests. - Annotated the unreachable `if (!this.client)` guard with a justified `v8 ignore` per the AGENTS.md dead-guard policy. Re-verified both regression directions: the `roots/list` assertion fails against `origin/v2/main`, the `rootsChange` one against `55e25b7c`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-3 fix — drop stale comments inside the roots handler (#1797) The four inline comments moved verbatim with the handler contradicted the method doc reworded one round earlier: they said "server's roots" (they are ours — it is a ClientNotification) and "we'll just dispatch an empty array", which is not what `this.roots || []` does when roots are configured. The new timing test asserts exactly that non-empty dispatch. Replaced with one accurate line. Comment-only. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(cli): advertise roots so roots/set is actually serviceable (#1797) Round-4 review found the same -32601 symptom on a second path: the CLI constructed its InspectorClient with no `roots` option, so `capabilities.roots` was never advertised and no `roots/list` handler was registered — yet `--method roots/set` calls `setRoots()`, which announces `notifications/roots/list_changed` to the server. A server taking up that invitation got "Method not found", after the CLI reported success. The review's suggested fix — register `roots/list` unconditionally — is not implementable: the SDK asserts the client capability inside `setRequestHandler` ("Client does not support roots capability"), so it throws during connect for any client built without the option, and `registerCapabilities` refuses to run after connect. A client that omits `roots` therefore cannot legally serve `roots/list` at all. Fixed at the layer that can fix it: `cli.ts` seeds `roots: []`, matching web (`App.tsx` always passes it). The core registration stays gated on the constructor value, now with the SDK constraint documented so the next reader doesn't retry the unconditional version. `setRoots()`'s doc claimed it "will enable roots capability if it wasn't already enabled" — it cannot, for the reasons above. Corrected, and dropped the dead `if (this.roots === undefined) this.roots = []` the very next line overwrote. Tests: a CLI test drives `list_roots` over an HTTP test server (verified load-bearing — without the seed it fails on the -32601 error text), and the core timing test gains a case pinning that roots set after connect are served with the current values, not the constructor's. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(cli,tui): advertise the roots configured in mcp.json (#1797) Round-5 review: seeding `roots: []` fixed the -32601 but not the outcome. `server-filesystem` discards an empty roots answer (`validatedRootDirs .length > 0`) and keeps its CLI-argument directories — which is the symptom this PR set out to fix. The user's roots were already in hand two lines away, on `serverSettings.roots` (lifted from mcp.json by `mcpConfigToServerEntries`). Now passed through `cleanRoots`, the same shared helper web uses, so both clients answer `roots/list` byte-identically for the same config. This matters most on the CLI because it is one-shot: `--method roots/set` sets roots on a connection torn down a few statements later, so the config file is the only durable way to give a run its roots. Same gap in the TUI (`App.tsx` built its options with no `roots` key at all), fixed the same way. The review suggested a follow-up issue for it; folding it in here instead, since it is the identical line and splitting it would leave a known-wrong client in tree. Also from review: hoisted `waiters` up with the other fields, and `injectRequest` now rejects after 1s naming the method and id rather than hanging to the vitest timeout. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-6 fixes — document roots in the CLI README, harden cleanRoots (#1797) - clients/cli/README.md enumerated the settings lifted from a catalog/config file without roots. Roots are the least discoverable member of that list — there is no flag, and `--method roots/set` dies with the connection — so the file is the only durable way to supply them. Said so. - cleanRoots() trusted its input's shape. `Root[]` is a compile-time type over hand-editable mcp.json, and as of this PR all three clients feed it straight from disk, so `"roots": [{"name":"Work"}]` or `"roots": "file:///work"` threw at connect. Non-array bails to []; an entry without a string uri is dropped with a warning. Hardened in the one shared place rather than three call sites. Review suggested a follow-up issue; folded in for the same reason as the TUI fix last round. - Test comment said the CLI passes a `roots: []` seed — true one commit ago. - injectRequest clears its timeout on reply (an armed timer outliving a test is the #1760 teardown class), and setRoots' doc drops the "despite what this comment used to claim" archaeology. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-7 fix — cleanRoots also throws on a non-string name (#1797) The round-6 hardening guarded `uri` and left `name` throwing one line later: `r.name?.trim()` guards null/undefined, not "is a string", so `"roots": [{"uri":"file:///a","name":42}]` still died at connect — the same crash class, one field over, reachable from all three clients and the web settings-save path. Reproduced against the shipped function body before fixing. A bad `name` now costs the name, not the root (it is optional, and the uri is still usable), with a warning. The doc no longer overstates what is validated. Also from review: the CLI test comment said "`cli.ts` now seeds `roots: []`" — true only of that ad-hoc test's path; and the registration comment's `(:588)` line reference is now a durable name. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): normalize roots in setRoots() too (#1797) Round-8 review: this PR made cleanRoots the single normalizer for roots at connect and at save, but setRoots — the third way roots enter the client — did a bare `this.roots = [...roots]`. Web cleans at its call site; the CLI does not, so `--method roots/set --roots-json '[{"name":"Work"}]'` stored a root with no uri, advertised it in the roots/list reply, and echoed it back as success. cleanRoots is idempotent, so calling it inside setRoots covers the CLI without disturbing web's pre-clean, and all three entry points now agree. Review filed this as an optional follow-up; folding it in for the same reason as the TUI and cleanRoots items — it is smaller than the issue describing it. Test asserts the malformed entry is dropped and the valid one survives; verified load-bearing against the bare-spread version. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-9 fixes — normalize roots in the constructor too (#1797) The round-8 commit claimed "all three entry points now agree"; review pointed out the constructor is a fourth, and the only one still raw. All three clients clean at their call site, so nothing was broken — but the invariant was enforced by convention rather than by the class, at the very entry point this whole thread started from. `this.roots` now runs through cleanRoots, with a ternary so the `undefined`-vs-`[]` distinction that gates capabilities.roots survives. Also from review: - The CLI roots/set test provoked a cleanRoots warning without suppressing it (AGENTS.md). Suppressed, and the spy now asserts the drop was reported rather than silent. - `enables roots when previously undefined …` in coverage-backfill was named after behaviour round 5 established does not exist. Renamed, with a comment on what that client actually can and cannot do. - setRoots dispatches a copy to rootsChange, as getRoots() already did, so a listener can't push into the list we advertise. - run-method.test.ts's connectStdio now passes `roots: []` like cli.ts, so the helper matches the client shape production builds. Verified load-bearing: the constructor test fails against the raw assignment. validate / coverage / build-gate / smoke / storybook all green (run as separate stages — full `npm run ci` kept being killed mid-coverage by the environment). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-10 fixes — copy at both rootsChange dispatches (#1797) Round-9 claimed setRoots' copy closed the "a listener can't push into the list we advertise" hole. It closed one of two sites: my reply conflated the notification handler's *dispatch* with the roots/list *request* handler — the latter is serialized to the wire immediately and is fine, the former hands listeners the internal array. Both copy now. Also from review: dropped the `({}) as never` in the new constructor test for the real factory the sibling tests use (this test never connects, so it is never invoked), and the needless `async` on it. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-11 doc fixes; merge v2/main (#1797) - cleanRoots' opening sentence still framed it as the settings-form normalizer, contradicting the paragraph below it that says every client feeds it from disk. It has four callers now; says so. - Marked the two runtime guards `Keep:` with the reason. The `Root[]` parameter narrows both to `never`, so "the compiler says this can't happen" is the argument someone will use to delete the checks rounds 6-7 installed. The reason now lives where the deletion would happen. - Noted the constructor-normalization case in the timing test's header, so it is discoverable from a file otherwise about the initialized race. Also merged origin/v2/main (#1796, loopback binding), which had advanced by one commit and touched the smoke scripts. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): clear queued peer requests when connect() fails (#1797) Round-12 review found a second-order consequence of the handler move. Before it, registration happened after every await in connect(), so the moment a peer request could be queued the connect could no longer fail — queue and failure path never overlapped. Registering before the handshake puts both the handshake and the post-connect `setLoggingLevel` inside that window. The catch sets status to error and drops the transport but never cleared pendingSamples/pendingElicitations; only disconnect() does, and connect() reaches it solely on the connect-timeout path (not the default). So a server that elicits at `initialized` and then fails the connect left a live pending-request modal on a dead connection — web derives it from the queue lengths with no status gate — and the elicitation's cancel() never fired, hanging any awaitUrlElicitation waiter. Extracted disconnect()'s clear-and-cancel into clearPendingPeerRequests() and called it from the catch, dispatching the change events so the UI drops the modal. CLI/TUI are unaffected either way (sample/elicit off). Test note: the first version of this test passed with and without the fix — its `requestedSchema` omitted `properties`, so the SDK rejected the inbound params and nothing was ever queued. Fixed the fixture; it now fails without the change with a queued ElicitationCreateMessage. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): clear the peer-request queue on a mid-session crash too (#1797) Round-13 review, two items. The test asserted only `getPendingElicitations()`/`getPendingSamples()`, which read the arrays directly — but the arrays are not what removes the modal. `usePendingClientRequests` tracks its own state off the pendingSamplesChange/pendingElicitationsChange events, so a regression that cleared without dispatching would strand a live modal on a dead connection and stay green. Both tests now assert the events fired with an empty payload; verified by deleting just the dispatches, which fails them. The connect-failure fix left the other way a connection ends uncovered: a mid-session crash goes through the transport's `onclose`, which sets status and fires `disconnect` but never touched the queues. Same two hazards, and arguably more reachable. `clearAndAnnouncePendingPeerRequests` wraps the clear plus its events, guarded on a non-empty queue so the paths that overlap (a connect failure whose dropCachedTransport also fires onclose) announce once; called from both. New test queues an elicitation, fires onclose, and asserts the queue and the event — fails without the change. Review flagged the crash path as a follow-up; folding it in, as with the earlier out-of-scope items — it is the same seam and three lines. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-14 fixes — make both teardown paths agree (#1797) - The crash path's comment claimed "cleared before the `disconnect` event so a consumer sees an empty queue" while disconnect() cleared *after* its own dispatch — an invariant stated as a property and held at one of two sites, the same shape as the round-10 finding. Moved disconnect()'s clear above its status/disconnect block so both paths agree; its change events stay batched with the other teardown dispatches. Full gate green, so nothing depended on the old ordering. - clearPendingPeerRequests' doc still named the `connect()` failure path as a caller; that path now goes through clearAndAnnouncePendingPeerRequests. Names the two real callers. - The mid-session test waited on `await Promise.resolve()` — the microtask-counting pattern this file replaced with injectRequest two rounds ago, working today only because the enqueue happens exactly one tick deep. Waits on the client's own `newPendingElicitation` instead. - The file header still called the constructor case "the last case"; it is third of five, and the two teardown cases went undescribed. Rewritten. - Restored the missing blank line between two tests, and noted why the sampling-event assertion holds for a queue that never held anything. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-15 fixes — scope the parity claim to the arrays (#1797) Round-14 made the two teardown paths agree on when the queue is cleared, and the comment then claimed parity flatly. It holds for the arrays but not the events: the crash path announces before `disconnect`, while disconnect() batches its change events with the rest of its teardown, so they land just after. Since the events are what drives the modal, the distinction matters to the next reader. Scoped the comment rather than un-batching — un-batching would trade a documented ordering for a real behaviour change with no gain. Same rule as round 14, one level down: check the other site, then check the other axis. Also gave the new `newPendingElicitation` wait the timeout that `injectRequest` documents, so a regression that stops the enqueue fails at the wait instead of hanging to the vitest timeout. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * test: pin disconnect()'s clear-before-announce ordering (#1797) Round-14 moved clearPendingPeerRequests() above disconnect()'s status block so a `disconnect` consumer sees an empty queue on both teardown paths, and round 15 documented it — but nothing enforced it. The existing `disconnect` listeners only count events, and the two teardown tests cover the crash and connect-failure paths, neither of which goes through disconnect(). So the one behavioural change in the stack was untested in both directions, and the tidy that undoes it (regrouping the clear with the batched change events) would have stayed green. Test records the queue length from inside a `disconnect` listener; the fixture's close() never fires onclose, so it exercises disconnect()'s own ordering rather than the crash path. Verified against the pre-round-14 position: `expected 1 to be +0`. Folded the newPendingElicitation wait into a shared helper now that two tests use it. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * test: pin disconnect()'s announce as well as its clear (#1797) Round-16 pinned the array half of disconnect()'s teardown and left the events unpinned — and nothing else in the suite covers them, so deleting disconnect()'s two change-event dispatches would have left a live modal after a user-initiated disconnect with the suite green. That is the path most likely to have a modal open when it runs. Asserted after the await rather than inside the disconnect listener, since disconnect() batches the events deliberately; verified by deleting the dispatches (`expected undefined to be +0`). Header rewritten by category rather than position: it has now gone stale twice in three rounds because it counted tests ("the last two") and enumerated paths that a new test falsified. It names the three teardown paths and what distinguishes them instead. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-18 doc fixes — scope the three-path claim, re-attribute cancel() (#1797) The header I rewrote last round claimed all three teardown paths clear before announcing. True only if "announcing" means the `disconnect` event: the connect-failure path emits none at all, and both it and the crash path clear after their `statusChange`. Scoped to what's true. Same shape as rounds 10/14/15 — a parity asserted in prose that holds on one of the axes it could mean — which is now the fourth time, in a sentence I wrote. ElicitationCreateMessage.cancel()'s doc still credited `disconnect()` teardown; this PR made clearPendingPeerRequests() its only caller, serving three paths. Re-attributed, and promoted the no-`onRemove` note to its own paragraph with the consequence spelled out — it quietly became a three-caller invariant during this PR, and "don't splice mid-iteration" reads like a style note until you know it would skip every other entry and leave those promises unsettled. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): gate roots/list on what was advertised, not on this.roots (#1797) Round-19 review found a latent wedge, reproduced before fixing: a client constructed without `roots` that later calls setRoots() has this.roots defined, so registerPeerRequestHandlers tried to register the handler on every subsequent connect() — and the SDK throws "Client does not support roots capability" from setRequestHandler when the capability was never advertised. That throw lands before the handshake, so the client can never reconnect. Worse than the -32601 this PR started from: the connection is gone rather than answering one method badly. Not a regression (the same gate lived after the handshake before) and unreachable from the three shipped clients, which all pass roots. But core/ is a library surface, and the failure is unrecoverable. Gate now reads a construction-time `rootsCapabilityAdvertised` — the thing the SDK actually asserts on — so this.roots stays free to change. setRoots' doc says what does happen on such a client: the roots are stored and readable, but no server can ask for them. Also scoped the header's claim that the connect-failure path emits no `disconnect`: true of connect()'s own catch, but on a real transport dropCachedTransport's close() fires onclose first, which does dispatch one — making the catch's own call the auth-recovery backstop. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-20 — derive the roots gate from the advertisement itself (#1797) Round-19's fix restored the invariant but not the derivation: rootsCapabilityAdvertised was computed from `options.roots` alongside the capability block rather than from it, so the two encodings of "did we advertise roots" could drift. Adding any term to the advertisement condition would leave the flag true with `capabilities.roots` absent — the same unrecoverable wedge, reintroduced silently and in the throwing direction. Now read off the built capability object after it is assembled, so divergence is structurally impossible rather than merely absent. Verified the reconnect test still fails against the old `this.roots` gate. Also: named the capability-lifetime category in the test header (the new test was none of the three it listed), moved that test below the teardown group so the group stays contiguous, and rewrapped the over-long comment line. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): gate elicitation/create on what was advertised too (#1797) Round-21 review asked round-20's question of the capability next door, and there the two derivations already disagreed — no future edit needed. `elicit` is typed `boolean | { form?, url? }`, so `{}` or `{ form: false, url: false }` is valid, enables no mode, and advertises no `capabilities.elicitation` — while the registration gated on `this.elicit` being truthy. Reproduced before fixing: CONNECT ERROR: Client does not support elicitation capability (required for elicitation/create) Since #1797 moved registration before the handshake, that throw means the client cannot connect at all. Same wedge as the roots one, one capability over, and reachable through a documented option value rather than a future edit. Gated on a new elicitationCapabilityAdvertised, derived from the built capability object exactly as the roots flag now is. Test constructs with `{ form: false, url: false }` and asserts connect() resolves; verified load-bearing. Also made the four construction-time capability inputs (sample, elicit, receiverTasks, advertisedExtensions) `readonly`. The round-19 bug was a mutator making a construction-time fact stale, so a future setter for any of these should be a compile error rather than a review catch. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): derive tasks.requests and the remaining gates from the advertisement (#1797) capabilities.tasks.requests declares which server→client requests we accept as tasks, but was built from `receiverTasks` alone — so `{ receiverTasks: true, elicit: false }` told the server it could send a task-augmented elicitation/create while advertising no elicitation capability and (since the last commit) registering no handler. A server taking the invitation gets -32601: the advertise-then-refuse shape this PR opened with, and the last place it survived. Now built from capabilities.sampling / capabilities.elicitation, both decided just above, with the key omitted entirely if neither applies. Also made the remaining three registration gates read the built capability object like the two from the last two rounds — sampling, tasks/*, and the URL-elicitation completion handler. None could diverge today, but a mixed convention means the next reader has to re-derive which gates are safe, and adding one condition to an advertisement block is exactly how elicitation broke. Five gates, one rule now. Test asserts a receiverTasks+elicit:false client advertises tasks.requests with sampling and without elicitation; verified load-bearing. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-23 — pin the omitted-requests arm, drop the count again (#1797) The `requests` key is omitted when neither sampling nor elicitation is advertised, but every `receiverTasks: true` construction in the tree left one of them enabled, so that arm was unexercised — and it would not have shown as a coverage failure, since one uncovered branch in a ~4900-line file sits well inside the per-file gate. Second client in the same test asserts `tasks` survives (list/cancel are still serviceable) while `requests` is absent rather than empty; verified against `requests: taskRequests`. The header regained a numeral four commits after round 17 rewrote it by category to stop exactly that, and it was already stale. Dropped, and named the third category — advertise-only-what-you-serve, which is about the capability object rather than a registration gate. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): settle queued samples on teardown instead of discarding them (#1797) Round-24 review. Round 13 recorded the samples-vs-elicitations asymmetry as deliberate on two grounds: nothing internal awaits a sample, and the transport is gone by then. The first still holds; the second is false on one path, and it is a path this PR created. `connect()`'s catch drops the transport only when `!transportHasAuthProvider`, so on the HTTP+OAuth path — web's ordinary authenticated case — it is retained and the caller reconnects over it. A sampling request queued during the handshake was then discarded with the connection still live: no response frame ever written, and the server waits forever for a reply to a request we accepted. Accept-then-silence, which is worse for the peer than the -32601 this PR opened with. SamplingCreateMessage.cancel() mirrors the elicitation one — settles (rejects, what the UI's decline path already sends and the SDK turns into an error response) without calling onRemove, since the caller iterates the queue and clears it. clearPendingPeerRequests loops it like the elicitations. Also corrected two comments that assumed the catch always tore the transport down, and noted that the retention is gated on transportHasAuthProvider alone — independently of isConnectAuthRecoveryError, which gates only the status hold. Test asserts settlement rather than a wire frame, since on the disconnect() path the transport is closed before the settle; a later respond() throwing "already resolved or rejected" is the proof. Verified load-bearing. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): mark receiver-task payload promises handled (#1797) Round-25 review. clearPendingPeerRequests iterates every entry in pendingSamples, and two shapes live there: the plain server-request one, whose reject reaches the SDK handler, and the task-augmented one, whose reject reaches the receiver task's payloadPromise. That promise only gets a consumer when the server polls tasks/result, so a task sitting in input_required has no handler attached — and the settle added last commit surfaced as an unhandled rejection. Not hypothetical: the repo already worked around it test-side, with a helper that reaches into private state to attach no-op catches, whose comment describes this exact race. Fixed at the source instead — one `void payloadPromise.catch(() => {})` at creation, which also covers the pre-existing cancelReceiverTask path. Removed the now-redundant helper and its two call sites. Verified: dropping the source line makes that suite report two unhandled rejections. Corrected SamplingCreateMessage.cancel()'s doc — its "no response frame is ever written" rationale is true of the plain shape only; a task-augmented sample already answered the wire with a CreateTaskResult, so settling it is about not leaving the task in input_required limbo. Nits: generalized the enqueue wait helper to both events so the sampling wait gets the timeout the elicitation one has; moved the sampling teardown test back into the contiguous teardown group; named the settle property in the header; fixed the fixture doc that still claimed to record replies. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-26 — drop the comment that outlived its call (#1797) The suppressReceiverPayloadRejections call went last commit; its two-line comment stayed, describing a pre-attach that no longer happens and crediting the test for handling a rejection the source now handles. Removed. Also reworded the surviving .catch() in inspectorClient.test.ts: it is still needed, but as a value capture for the assertion, not for unhandled-rejection suppression. Rewrapped the teardown paragraph in the test header — my round-25 edit left a 104-char line, and two hand-fixes left it ragged, so it is reflowed in one pass. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-27 — skip the roots/list_changed the SDK would refuse (#1797) Review reported setRoots() emitting roots/list_changed on a client that never advertised roots, inviting a re-fetch we'd answer -32601. Probed it: the premise is wrong. The SDK asserts the notification capability too — `notification()` rejects with "Client does not support roots list changed notifications" and nothing reaches the wire. No spec violation and no invitation; what actually happened is that setRoots provoked that rejection and logged it as "Failed to send", implying a transient failure where the truth is structural. Kept the early return on that corrected basis: it stops us provoking a rejection we know will come, and warns accurately instead. Doc and comment say the SDK is the enforcer rather than claiming this guard is what keeps the wire clean. The test pins the end state (no invitation for roots we can't serve) and passes either way — recorded as such rather than presented as a regression guard, since it isn't one. Also moved the advertise-category test below the two gate cases, matching the order the header introduces them. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-28 — gate the notification on the predicate the SDK asserts (#1797) The roots/list_changed guard read `capabilities.roots !== undefined`, but the SDK asserts `capabilities.roots.listChanged` for the notification — wider than the flag. Identical today (roots are only ever advertised as `{ listChanged: true }`), but a conditional advertisement would un-gate it silently, which is the round-21 shape again. Split into rootsListChangedCapabilityAdvertised, mirroring urlElicitationCapabilityAdvertised, which already reads its assertion's sub-field; rootsCapabilityAdvertised keeps the roots/list registration, where presence is what setRequestHandler checks. Also moved the new test out of the gate pair it had split — it is a runtime-emission case, not a registration-gate one — and extended the header's converse-category paragraph to cover notifications, which indexes it without inventing a category. Reflowed the setRoots doc paragraph in one pass (111-char line from the last commit, plus a ragged orphan from an earlier edit). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-29 — chase the corrected premise into clients/cli (#1797) Round 27 corrected "setRoots announces roots/list_changed to a server that then gets -32601" — the SDK refuses the notification from a client that never declared it, so nothing reached the wire. That reword happened in core/, but the same premise had been written into the two comments justifying the CLI's roots seed, and they still asserted it. Both now state the real pre-fix failure: a server asking for roots on its own got -32601, and roots/set could not announce at all. Also fixed the warn message under the round-28 gate: it named the roots capability while the gate is on roots.listChanged — precisely the pair that split was made to distinguish, so a client advertising `{}` would have been told it advertised nothing. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): reject in-flight raw-wire requests on a mid-session crash (#1797) rejectPendingRawWireRequests had one caller — disconnect() — so the crash path settled the inbound peer-request queue this PR added and left the outbound raw-wire map alone. That map holds the modern tasks/* frames the SDK's era gate refuses to route, so the SDK's chained _onclose doesn't know about them: a Tasks-tab poll in flight when the server died waited out its own 30s timeout and reported a timeout for what was a crash, on a connection that had already dispatched `disconnect`. Pre-existing — both the map and the crash path predate this PR. Fixing it here because the invariant it violates is one this PR wrote down: "every path that ends a connection has to clear that queue, announce it, and settle each entry", in clearAndAnnouncePendingPeerRequests' doc and the timing test's header. One channel was true of one path out of three. Test starts a raw-wire request and fires onclose; without the call it hangs to the vitest timeout, which is the symptom. Checked the two siblings and left both: taskInputAbortControllers is already unblocked via the elicitation cancel the crash path performs, and activeToolCallAbortController rides an ordinary client.request the SDK settles itself. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-31 test fixes — justify the cast, drop the no-op tick (#1797) - The new test's `as unknown as` had no justification, which AGENTS.md explicitly prohibits. It is unavoidable — rawWireRequest is private and every public route is gated on isTasksExtensionNegotiated(), which needs a modern handshake the fixture doesn't perform — so that reasoning is now written down instead of left to be re-derived. - `await Promise.resolve()` was a no-op: rawWireRequest registers its entry in the promise executor, synchronously, before any suspension. Dropped, which also makes the test pin that ordering rather than paper over it — and stops the third recurrence of the microtask-counting pattern this file's own helper doc argues against. - Moved the test into the teardown group and extended the header's teardown paragraph to cover the outbound direction, which it framed entirely as inbound. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-32 — separate the outbound teardown claim from the inbound one (#1797) My round-31 header edit spliced the outbound raw-wire passage into the middle of the inbound claim, so "There are three: a failed connect(), a mid-session transport close, and an explicit disconnect()" ended up reading as scoping the outbound settle — which is two by design, since nothing populates the map before the handshake. Round 31 established that scoping so the code wouldn't imply an impossible state; the prose then un-scoped it. Split into its own paragraph with its own scope clause, and reflowed both in one pass rather than nudging breaks. Also raced the raw-wire assertion, for the reason the file's own wait helper documents: without the teardown rejection the test waited out the request timeout and reported a bare vitest timeout. It now fails with "raw-wire request not settled by teardown" — verified. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-33 — extract the timeout race, mark the failure-path promise handled (#1797) Three copies of "reject after 1s so a regression fails where it happened" had accumulated, each re-explaining itself. Extracted `withTimeout`, used by the enqueue wait and the raw-wire assertion; injectRequest keeps its own since its timeout also deletes the waiter entry. On the failing path the raw-wire test left the production 30s timer armed, so a regressed run got the intended failure plus a stray unhandled rejection up to 30s later — the armed-timer class the file's own comment names, on the failure path. `settled` is marked handled before the race. Verified: with the fix removed the run fails with "raw-wire request not settled by teardown" and no unhandled rejection. Recorded why disconnect() drains the SDK's in-flight requests but rejects the raw-wire map outright: the user asked to disconnect, and a Tasks poll is not worth delaying that for. Pre-existing and deliberate, now written down where the asymmetry is visible. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-34 — drop a no-op guard and the false mechanism it asserted (#1797) Round 33 claimed the raw-wire test could leave `settled` rejecting unhandled on a regression run, and round 33's fix encoded that as `void settled.catch(() => {})` with a comment explaining it. Both were wrong: `Promise.race` attaches a rejection handler to every input, so `withTimeout(settled, …)` on the next line already marks it handled. Verified with a standalone script — no unhandledRejection fires, and the pre-fix version was equally safe. Removed the line and the comment. The comment mattered more than the line: it told the next reader that `Promise.race` does not mark its losers handled, which is false and would cost them elsewhere. Kept the true half by setting `timeout: 50` on that test's client, so a regression leaves no 30s timer armed and fails with `Raw request "tasks/get" timed out after 50 ms` — more self-describing than the raced generic message. Verified. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-35 — the drain comment described a trade-off nobody made (#1797) Last round's comment said raw-wire requests "get no share of the drain above: the user asked to disconnect, and a Tasks poll is not worth delaying that for" — framing a decision. Verified it isn't one: `safeDisconnectTimeout` defaults to 0 and the only caller anywhere that passes a value is a single test, so nothing is drained in production; and the drain polls the SDK's `_responseHandlers`, which never holds raw-wire ids, so those were never eligible regardless. Reworded to what is true. The framing came from my own round-33 reply, which described the drain as running on every disconnect — a comment faithfully encoding a half-read mechanism, which is the same failure as the `Promise.race` one it followed. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-36 — the race comment outlived the mechanism its own commit replaced (#1797) `timeout: 50` made the request settle in ~50ms, so the 1s race can never win and a regression fails on the assertion (`timed out after 50 ms` vs the expected /Connection closed/) — which the commit message said, while the comment beside it still credited the race. Reworded: the short timeout is what fails fast and names itself, the race is the backstop for a promise that never settles at all. Race kept; it covers the case where the request's own timer is broken too. Third comment in three rounds encoding a mechanism the same commit changed. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-37 — scope the timeout comment, align the bypass guards (#1797) `timeout: 50` is the client-wide request timeout, not a bound on the raw-wire request; it is inert in that test only because the test issues no SDK request after connect. Noted, so someone adding one later knows it would inherit a 50ms budget rather than reading the comment and looking elsewhere. The two installReceiverTaskResponseBypass guards read `this.receiverTasks` while the five gates around them read the advertisement. Installing the bypass mutates the Protocol's _requestHandlers map, so it is registration — same as the setRequestHandler it follows. Equivalent today; aligned so the convention is uniform rather than something the next reader has to re-derive per site. The genuinely-runtime `receiverTasks` reads are left. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-38 — one predicate for the whole bypass mechanism (#1797) Round 37 moved the bypass *install* guard to the advertisement and left the three branches it controls on the raw option, splitting one immutable fact across two sources. The dead direction is the harmful one: if the tasks advertisement ever acquires a condition — the exact edit that produced the elicitation wedge — then advertised goes false while receiverTasks stays true, the bypass is not installed, and the handler still returns `{ task }` into the SDK's validating wrapper, which rejects it. That is the breakage the bypass exists to prevent, and before round 37 it could not happen because install and branch read the same field. All three now read tasksCapabilityAdvertised. `receiverTasks` is left where it belongs: assigned once, and driving the advertisement. Also noted that the wrapper's own check is redundant with its install guard and kept deliberately, since it must agree with the handler branch. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): don't carry receiver tasks into the next session (#1797) Round-39 review found the last member of the enumeration rounds 12-14 and 31 built: receiverTaskRecords was cleared only by disconnect(). The crash path doesn't clear it, the connect() failure catch doesn't, and connect() never reset it — so a record created in one session survived into the next connect() on the same instance, and `tasks/list` is answered from that map. A new server would be told about a task it never created. Reachable in web (the only client with receiverTasks): a task-augmented sampling/createMessage arriving in the handshake window creates a record; if that connect then fails with a 401, App.tsx retries connect() on the same instance. Cleared at the top of connect() rather than only on the crash path — the auth-recovery retry means ending a session isn't the only way a new one begins, and starting clean covers every route in. disconnect() keeps its clear via the same extracted helper. Pre-existing — the crash path never cleared these — but the handshake window this PR opened is what lets a receiver task exist before connect resolves, and this was the one member left outside the invariant the PR documented. Test verified load-bearing. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): reset all session-scoped state on connect, not just receiver tasks (#1797) Round-40 review: receiverTaskRecords was one of five collections cleared only by disconnect(). Two of the other four have real symptoms on the same crash-then-reconnect path: - subscribedResources — the modern subscribeToResource early-returns on a uri already in the set, so after a reconnect the user's Subscribe click silently sends nothing to the new server while the UI keeps listing it. - cancelledTaskIds — a leftover id mislabels a *new* task sharing it as cancelled rather than failed, and server task ids are often short. taskInputAbortControllers leaked un-aborted controllers. modernLogLevel had the opposite asymmetry: disconnect() cleared it so it couldn't leak (#1629), but nothing restored it, so a reconnect after an explicit disconnect silently lost the user's configured level while a reconnect after a crash kept a mid-session override. It is now recomputed from serverSettings the way the constructor does, so both routes agree. clearReceiverTasks became resetSessionState, called from the same place. Tests cover the two symptomatic siblings and now assert receiver tasks through the `tasks/list` handler — what a server actually sees, and the symptom itself — which also removed the unjustified `as unknown as` the previous commit added. The remaining cast is justified in place, since cancelledTaskIds has no public reader. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(web): re-seed the Logs control when the client re-seeds its own level (#1797) Round-41 review caught a consequence of the last commit. `resetSessionState` restores the client's modern log level from the server setting at connect; `resetSessionScopedUiState` still blanked the web control to null on every disconnect. On the one path that reconnects the *same* instance — auth recovery — the control then read Off while every modern request carried `_meta` logLevel "debug", the default. A tool whose job is showing what you send should not disagree with itself about what it sent. The control now re-seeds from the active server's settings exactly as setupClientForServer does, instead of blanking. Also from review: - Added the tests the restore never had, covering both halves: a mid-session override does not survive a reconnect, and a disconnect no longer drops the configured level (the #1629 half, and the actual behaviour change). - resetSessionState's doc claimed "every piece of state scoped to one connection", which invites assuming membership rather than checking it. Now states the rule: state a new session would misread; what only needs settling on the way out stays with the teardown paths. - Header: session-scoping paragraph moved below the outbound one to match test order, and it now names all five members rather than three. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * refactor(core): one derivation for the modern log level (#1797) Round-42 review: the configured level was computed independently in four places — twice in InspectorClient (constructor, resetSessionState) and twice in App.tsx (the seed and the re-seed). The client and the Logs control agreed only because four sites happened to compute the same expression, and any drift reproduces exactly the symptom the previous commit fixed: the control reading one level while every modern request stamps another. Extracted resolveModernLogLevel into core/mcp/types.ts beside DEFAULT_MODERN_LOG_LEVEL; all four call sites use it, with the web mapping undefined→null at its own boundary (display, not a second derivation). Also from review: - disconnect()'s comment still credited itself with preventing the #1629 leak, which resetSessionState now prevents. It says what the line does now — read "not opted in" while disconnected — and notes the web control deliberately shows the configured level in that window. - resetSessionState's doc claimed an un-aborted taskInputAbortControllers entry "leaks a paused poll loop". Both registration sites release in a finally, so nothing leaks permanently; the abort closes the window between a crash and the unwind. Softened, and it now has the test it lacked — the only member of the five without one. - Noted in App.tsx that the re-seed depends on activeServerIdRef still holding the outgoing id (it syncs in a passive effect), so an eager clear would silently drop it to the default. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): reset the subscription stream state with the set it derives from (#1797) Round-43 review: resetSessionState cleared subscribedResources but left modernStreamState, so a crash-then-reconnect on the same instance held an empty set alongside `{ active: true }` — a combination the rest of the file treats as impossible, since onModernSubscriptionClosed and onModernReconnectFailed both compute `active` from `subscribedResources.size > 0`, and the comment at :4712 asserts a terminal drop clears both. No symptom today only because the UI mirror resets itself independently, in another file — which is the arrangement this PR has spent its rounds removing. Clearing it here also announces the reset; that set's every other mutation dispatches. Also from review: - The web re-seed treated "no server found" as the default level rather than Off, because resolveModernLogLevel returns the default for absent *settings*. Reachable via a second disconnect after a crash, and via removing the active server. Now distinguishes the two. - Converged the two injectRequest copies: withTimeout takes an optional onTimeout, so both drop their waiter on the timeout path. The newer copy had lost exactly the cleanup that was round 33's reason for leaving the original alone. Test seeds the stream state a live subscription would have left; verified load-bearing (the earlier version passed either way, since the fixture never activates a stream). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(web,core): correct the re-seed branch; share the subscription-stream reset (#1797) Round-44 review. My round-43 fix traded one wrong branch for another: it read `?.settings`, which is undefined both when no server matched *and* when the server has no settings node — and the second is the common case (a hand-written mcp.json entry never opened in Server Settings). Those servers showed Debug while connected and Off after disconnect, which is exactly the client/UI disagreement rounds 41-42 removed, with no modernLogLevelChange event to reconcile it. Now branches on the server. Worth noting no test would have caught it: App.tsx is a documented exclusion from the coverage gate. Also from review: - resetSessionState cleared subscribedResources without announcing it, two lines below a comment arguing that every other mutation of that set dispatches. And it took two statements of a six-statement cluster disconnect() clears together — leaving modernReconnectAttempts (a new session's first drop would back off as if it were the old one's nth), modernListenGeneration (the bump that makes an in-flight re-listen bail), and a stale modernSubscription. One resetSubscriptionStream helper now serves both, announcing both axes; disconnect() keeps only the live stream's best-effort close. That also removed the file's one assignment to modernStreamState that bypassed its setter. - The stream-state test pinned the value, not the dispatch — the half the UI actually tracks. Now asserts both; verified against a bare field assignment. - onModernSubscriptionClosed's comment claimed the disconnect reset always clears the set; on a crash it doesn't until the next connect. Says so. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-45 — reattach the orphaned doc, order the announcements (#1797) Extracting resetSubscriptionStream last round inserted it *and its doc* between resetSessionState's doc and resetSessionState — so two doc blocks stacked, the new helper took the nearest, and the 23-line block explaining each member's symptom attached to nothing. Dead text where it sat, and the most load-bearing comment in the batch. Moved, not reworded. That is round 32's finding in production code, from the same cause: inserting anything between a doc block and its method is a move that needs checking, prose or code. Also from review: - resetSubscriptionStream announced the cleared set before updating the stream state derived from it, so a listener in between saw an empty set with an `active` stream — the combination the helper exists to prevent, exposed by its own dispatches. Both fields move first, then both are announced. The onModernSubscriptionClosed comment now says that window is never observable, which is newly true. - The test header called the log level "cleared"; it is re-seeded, which is the #1629 half of the test twenty lines below. Says "reset", with the distinction spelled out, and the paragraph reflowed in one pass. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * fix(core): close the listen stream the connect-path reset drops (#1797) Round-46 review: resetSubscriptionStream's doc claimed disconnect() is the only caller with a stream worth closing, and the helper did a bare `modernSubscription = null`. Not guaranteed on the connect path — an `onerror` without an `onclose` leaves the transport up, connect() reuses it, and the reset then drops the last reference to a stream still open on the server. Nothing can close it afterwards: the `closed` handler bails on the bumped generation, and `closed` never resolves for a live stream. Before the extraction the surviving reference was eventually closed by refreshModernSubscription as `previous`, so this was newly unclosable. The helper now closes it itself, best-effort, after the dispatches so the round-45 ordering is unaffected; disconnect() drops its two hand-rolled lines. Also from review: - The session test pinned the subscription set's value but not its dispatch — the half the UI tracks, and the gap round 43 closed for the stream axis one line away. Both now asserted; verified load-bearing. - resetSessionState's doc said disconnect() "clears all of this", which is true by duplication for two of the five members; it now says a sixth has to be added in both places. And "the teardown paths" handle the in-flight tool call only on the disconnect route — on a crash it is callTool's own finally. Named. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-47 — fix the doc's count, catch the closed-promise chain (#1797) - resetSessionState's doc said "two members through the same helpers, the other two hand-rolled" for a body with five. It is two and three — and the member the enumeration dropped is the one that is *paired* rather than duplicated: modernLogLevel is re-derived here and blanked in disconnect(), deliberately. Both corrected. - `subscription.closed.then(...)` had no rejection handler, unlike every other promise around that stream. Newly reachable: before the last commit a connect-path reset dropped the reference with `closed` permanently pending, so the chain never settled at all; closing the stream is what makes it settle. Added, with the Node-terminates reasoning the file already applies to payloadPromise. - The listen-stream test exercised transport reuse without asserting it — the half that makes "still open on the server" true. Counts factory calls now, so the premise fails loudly if a future change recreates the transport rather than silently ceasing to apply. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-48 — scope the catch, guard the close, index the test (#1797) - The .catch comment claimed the connect-path close made a *rejecting* `closed` reachable. It doesn't: closing resolves it. What that close newly reaches is the handler running at all, where the reference used to be dropped with `closed` pending forever. And chaining after `.then` also swallowed handler throws — silently abandoning a re-listen on the path whose job is recovering a dropped stream. Now the second argument to `.then`, scoped to the rejection, with the comment saying which half is which. - `closing?.close()` reaches into third-party code from `disconnect()`'s straight-line teardown, none of which is inside a `try` — a synchronous throw would skip the raw-wire rejects, the task aborts, the receiver-task clear and nine dispatches, silently, since most callers catch a disconnect(). Best-effort now against both failure modes. - Header indexes the listen-stream test: same category, but a release obligation rather than a misread one. - Its comment described the generation guard, which the stub can't exercise; marked as production behaviour. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011TrCvpunxToe7qnppk6zwr * review: round-49 — pin the best-effort close, unorphan the doc (#1797) Round 48 made `resetSubscriptionStream`'s close best-effort against both a rejected promise and a synchronous throw, but nothing exercised either arm: both `modernSubscription` stubs in the tree resolve. On the `disconnect()` path an escaping failure would silently skip the raw-wire rejects, the paused task-input aborts, the tool-call abort and `clearReceiverTasks()` — the findings rounds 13 and 30 closed, re-opened at once with a green suite. Two table-driven cases now seed a `close()` that throws and one that rejects, and assert teardown continued past it. Also: the `.then` comment gave opposite verdicts on one mechanism (an unhandled rejection is named unacceptable, then chosen for a handler throw) — reworded to say the handler cannot throw today and why the shape is defensive; and `refreshModernSubscription`'s doc block, orphaned two methods up since before this branch, moved onto the method it describes. Closes #1797 * review: round-50 — sweep the queues start-clean, unify the close (#1797) `resetSessionState`'s doc excluded the peer-request queues and the raw-wire map because "the teardown paths handle them" — true of the paths that run. An `onerror` without an `onclose` runs none of them: it flips status to "error" and leaves `baseTransport` cached, so a `connect()` on the same instance reuses a live transport. That is the route the round-46 stream close was built on, and it strands these two the same way — the web pending-request modal is derived from the queue length with no status gate, so it outlives the session, and answering it writes a response for the old request id onto the new connection. Both are now swept start-clean beside the reset (both helpers are idempotent, so the routes that already ran them are unaffected), the end-clean clears stay put for the `disconnect`-event consumer, and both docs say which is which. Also folds the three subscription `close()` sites onto one `closeSubscriptionBestEffort` helper, so all three absorb a synchronous throw as well as a rejection — at `refreshModernSubscription`'s first site the reference is already dropped, so a throw there both leaks a live stream and aborts the refresh before its replacement is opened. Closes #1797 * review: round-51 — pin the sweep's ordering, name the category (#1797) The connect-start sweep must sit after `resetSessionState()`, and the two calls read as independent. Cancelling a task-augmented peer request settles it synchronously into the record callback, which ends in `upsertReceiverTask` — a no-op only because `clearReceiverTasks()` just emptied the map. Hoisted, it would emit a `notifications/tasks/status` for the outgoing session's task onto the transport the connect is about to reuse. Stated at the call and pinned by a test (which fails on the hoist). Adding that third caller also made five docs stale, all describing the old set by count or as "teardown" — the class round 23 diagnosed. All five now name the category: every route out, plus the top of `connect()` as a backstop for the one route in that settles nothing. And the site where round 50's fold actually changed behaviour — `refreshModernSubscription`'s first close, where the reference is already dropped and a `.catch()` never caught a synchronous throw — is now tested from the public surface (a re-list over a poisoned stream), table-driven over both arms. Verified to fail on the pre-fold form. Closes #1797 * review: round-52 — say what the sweep prevents, scope the close doc (#1797) The sweep's comment named as its harm the exact write the sweep performs: on this route the SDK's transport is still live, so cancelling a queued peer request emits a response for the old id immediately — which is right, and is the settle-don't-discard rule, at the best available moment. As written it read as an argument against emitting the frame, inviting the clear-without- settling regression rounds 24–25 removed. It now names the real hazard: a *user*-authored answer landing arbitrarily later, past the re-handshake. `closeSubscriptionBestEffort`'s "at each site" held at two of three — the superseded-generation discard stored nothing and has nothing following it, so its escaping failure would only reject an already-superseded `subscribeToResource`. Scoped, still wrapped. Also moves the `disconnect()`-side close pair next to its `connect()`-side twin, and states which axis the file's ordering follows, since route and category both describe it and cut differently. Closes #1797 * review: round-53 — give each close site its real harm, pin it (#1797) The doc grouped `resetSubscriptionStream`'s close with the "takes the caller's remaining work with it" sites. It can't: the call is `void`-ed onto an `async` helper, so even a synchronous `close()` throw becomes a dropped rejection — `disconnect()`'s teardown runs regardless. The harm there is an *unhandled* rejection, which this PR already established (the payloadPromise fix) is the process-fatal class in Node — so the one site with the severe failure mode was documented as having the mild one. Each site now carries its own consequence, and the local comment no longer gives two reasons that cancel each other. The test had the same gap: both its assertions hold with the helper's try/catch deleted. It now also asserts no unhandled rejection fired, which does fail without the guard (verified on both arms); the teardown assertions stay, since they still catch unwrapping the call to a bare `close()`. Also names the two back-references the round-52 move left positional, and softens the "on the wire until dropCachedTransport()" claim — that drop is conditional and never runs on stdio. Closes #1797 * review: round-54 — the mildest site isn't mild from every caller (#1797) `refreshModernSubscription` has three callers, and the superseded-generation discard's failure reaches all of them. From `unsubscribeFromResource` — which deliberately keeps the removal when a re-listen fails — it surfaces "Failed to unsubscribe" for an unsubscribe that actually stuck, which is worse than the consequence the doc assigned to the site it called mildest. The re-listen bullet had the mirror gap: on the reconnect caller the escape is absorbed into the backoff run, not a set left with no stream. Both now name their callers, and the generation-bump claim is softened to what it proves (a newer refresh *started*). Test: the `unhandledRejection` listener is process-wide, so the assertion now filters to this fixture's own reason rather than failing on a rejection some other test left pending; and it is registered inside the `try` whose `finally` removes it, so a throw in the setup can't leak it into the rest of the file. Re-verified the filtered form still fails on both arms with the helper's try/catch removed. Closes #1797 * fix(core): don't roll back a subscribe a newer refresh superseded (#1797) `subscribeToResource`'s rollback assumes what its comment says — that the server filter is unchanged — which is exactly what supersession falsifies. The newer refresh built its filter from the set *including* this URI, so on success the server honors it; rolling back leaves the set missing a URI the live stream carries (the UI showing it unsubscribed while its resources/updated keep arriving) and, if it was the only one, an empty set with an active stream — the combination `resetSubscriptionStream` exists to prevent, reached from the other end. Reachable without any close() failure: a rejecting `listen()` propagates the same way, which a user subscribing while the reconnect timer re-lists — i.e. exactly when the server is flaky — can produce. The rollback is now gated on the generation not having advanced past this call's own bump; the error still reaches the caller either way. Test drives two overlapping subscribes with the first `listen()` rejecting, and fails without the gate. The doc bullet that called this the mild site is corrected with it: both user-initiated callers report a failure that did not happen, in mirrored ways. Also gives the two close-failure arms distinct reasons so the unhandled-rejection filter can exclude the sibling arm as well as a foreign rejection. Closes #1797 * fix(core): reconcile the stream badge when a re-listen this call owned fails The supersession gate left the both-fail case stuck: two overlapping subscribes whose listens both reject leave the superseded URI in the set (its rollback correctly skipped) while the state keeps the optimistic "connecting" — no stream, no reconnect armed (neither onModernSubscriptionClosed nor onModernReconnectFailed ran), so the badge never changes and no resources/updated ever arrive. That is the caveat this file already wrote down for the neighbouring case — the bump proves a newer refresh *started* — applied to the gate itself. Fixed in the branch that knows the answer: the call that still owns the re-listen reconciles the state instead of only handling an emptied set, reusing the "ended" badge onModernSubscriptionClosed already reports for "stream gone, URIs still subscribed". unsubscribeFromResource gets the same treatment — it rolls nothing back by design, but its failed re-listen left the badge reading Listening over no stream. Also: the gate's comment named one of the two things that bump the generation (resetSubscriptionStream also does, from disconnect and the start-clean reset) and cross-referenced the wrong invariant; both corrected, and `<=` tightened to `===`, which is what the monotonic synchronous bump actually guarantees. Closes #1797 * fix(core): order the subscription writes, retry instead of stranding Two sites emptied the subscription set and announced it before the stream state caught up — the last-URI rollback in subscribeToResource's catch, and (pre-existing) the ordinary last-URI unsubscribe, which dispatches a round-trip before the re-listen sets INACTIVE. Both exposed an empty set with an active stream: the pair resetSubscriptionStream orders its own writes to prevent, and which three comments in this branch now cite. State before the announce at both, asserted from the consumer's side. And the reconcile added last round settled for an honest but dead badge: it was the only route to "stream gone, URIs live" that had made no attempt at all, over subscriptions the server may never have honored, recoverable only by an Unsubscribe/Subscribe toggle (a fresh Subscribe early-returns). It now hands them to the reconnect machinery, which already fits — base delay after a user-initiated refresh, bails on a terminal status or an emptied set, and lands on the same "ended" badge past the cap. That also removes the provably-dead null check the reconcile carried. Closes #1797 --------- Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>	2026-07-27 11:32:23 -04:00
Cliff Hall	ec185f1053	Logging: per-request log level on modern servers (era fork) (#1629) (#1724) ... * Logging: per-request log level on modern servers (era fork) (#1629) On modern (2026-07-28) servers `logging/setLevel` and the session-scoped level are gone: the client opts into logs per request via the `io.modelcontextprotocol/logLevel` `_meta` key, and the SDK does not attach it automatically, so server logs are silently absent until stamped. Core (`InspectorClient`): - `setModernLogLevel(level | undefined)` / `getModernLogLevel()` store the opt-in level; `mergeMeta` stamps it on every outgoing request, gated on the negotiated modern era so legacy connections are untouched. - Cleared on disconnect so it can't leak into the next connection. Web UI: - `LogControls` forks on `protocolEra`: legacy keeps the `logging/setLevel` selector + Set button; modern shows a per-request opt-in Select (Off + levels, applied immediately) with copy explaining logs arrive on the originating request's stream. - Threaded `protocolEra` / `modernLogLevel` / `onSetModernLogLevel` through LoggingScreen and InspectorView from App. Tests: unit coverage for the modern LogControls/LoggingScreen fork; integration coverage proving the `logLevel` `_meta` is stamped on a real modern server once opted in (and not on legacy). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * test-servers: logging era-fork showcase configs + proof screenshots (#1629) Add two manual-testing configs and README docs for the logging era fork: - logging-legacy-http.json: plain streamable-HTTP server with `logging: true` and a `send_notification` tool. The Logs tab shows the legacy session-scoped "Set Active Level" selector + Set button; calling the tool streams logs. - logging-modern-http.json: same, with `transport.modern: true`. Connect with Protocol Era = Modern and the Logs tab shows the per-request opt-in control; the client stamps `_meta["io.modelcontextprotocol/logLevel"]` on every request (visible in the Network tab request body). pr-screenshots/ documents the verified behavior end-to-end against both real servers: the legacy control with logs streaming, the modern per-request control, and the modern tools/call request body carrying the stamped logLevel _meta. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * test-servers: send_notification emits via per-request notifier so modern logs stream (#1629) The `send_notification` tool sent its `notifications/message` through the global `server.server.notification()`. On the modern (2026-07-28) stateless leg that isn't tied to the in-flight request, so the SDK answered the tool call with plain JSON and dropped the log — nothing reached the Inspector's Logs panel. Prefer the per-request `extra.sendNotification` (rebuilt from `ctx.mcpReq.notify`): scoped to the current request, it makes the SDK upgrade the response to SSE and stream the log back on the originating request's stream. Falls back to the global notifier (legacy path / no per-request context). Verified end-to-end: opting in via the Logs-tab "Log Level per Request" control and calling send_notification now streams the entry on the modern server; legacy behavior unchanged (server-notification integration tests green). README documents the modern logging showcase and the caveat that this test server streams whenever a tool emits — the SDK beta strips the per-request envelope from the handler context, so the tool can't gate on the client's logLevel opt-in (the client-side `_meta` stamp #1629 owns is still fully visible in the Network tab). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * test-servers: send_notification uses SDK request-scoped log() → spec-faithful modern gating (#1629) The previous fix routed `send_notification` through `extra.sendNotification` (`ctx.mcpReq.notify`), which streams on the modern leg but emits unconditionally — the log showed up even when the client hadn't opted in via the per-request `logLevel`. The SDK already provides the correct primitive: `ctx.mcpReq.log(level, data, logger)` (added in `McpServer.buildContext`). It applies the era-correct threshold gating itself — on the modern (2026-07-28) leg it reads the per-request `logLevel` opt-in from the request envelope and drops the message when the client didn't opt in or the level is below the requested severity; on legacy it honors the session level from `logging/setLevel` — and emits through `notify`, so the modern response still upgrades to SSE and the log rides the originating request's stream. Surface it as `extra.log` (`toHandlerExtra` maps `ctx.mcpReq.log`) and have `send_notification` prefer it, falling back to the global notifier for callers without per-request context. Verified end to end: opting in via the Logs-tab "Log Level per Request" control stamps `logLevel` in the request `_meta` and streams the log; setting it back to Off omits the key and the server gates the log (plain JSON response, no notification). Legacy server-notification integration tests still green. (No SDK change needed — the earlier apparent gap was our test server using the wrong low-level API.) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * web+core: add "Log Level per Request" as a per-server setting, defaulting to on (#1629) Adds the modern per-request log level as a persisted Server Settings option (Options → "Log Level per Request", below Protocol Era), so a server can opt into logs by default instead of the Logs-tab control starting at Off. - types.ts: `ModernLogLevel` (`"off"` | the 8 levels), `DEFAULT_MODERN_LOG_LEVEL` = "debug" (default-on so a modern connection surfaces server logs out of the box), `MODERN_LOG_LEVELS`, and an `isModernLogLevel` guard. Field added to `InspectorServerSettings` and `StoredMCPServer`. - serverList.ts: round-trip the field (lift on read, omit-on-default on write, drop unknown literals) mirroring `protocolEra`; add to the persisted field allowlist. - remote/node/server.ts: validateSettings accepts/validates it and the normalize path drops a malformed literal. - inspectorClient.ts: the constructor seeds `modernLogLevel` from the setting (`"off"` → not opted in), so a modern connection stamps the level from the first request without UI interaction. mergeMeta still gates on the modern era. - ServerSettingsForm/Modal: a Select (Off + 8 levels) with a modern-era description. App seeds the live Logs-tab control from the setting on connect so the two stay in sync. Tests: form Select (default/onChange/off/reflect), modal handler, serverList round-trips, route validation (accept/reject/drop), and constructor seeding (seeded level stamped from the first request; "off" opts out). The prior modern-era "stamps once set" test is updated for the new default-on behavior. Coverage gate green. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * Fix CI + review: format ServerSettingsForm, drop stray network export, broaden ignore - Prettier-format ServerSettingsForm.tsx (the format:check failure that broke CI — my pre-push `npm run format` from the repo root was a silent no-op since there is no root format script; formatting is per-client). - Remove `inspector-network-logging-modern-2026-07-19T15-11-27.013Z.json`, a Network-tab export debugging artifact accidentally committed in ca2aa526 (flagged by review). - Broaden the .gitignore debugging-artifacts section with `/inspector-network-*.json` so future network exports are ignored (the existing `/inspector-history-*.json` didn't match the network export naming). Verified with the full `npm run ci` (validate → coverage → smoke → storybook), green across all clients. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * test: add modern-era log delivery + gating e2e tests (#1629) The modern-era suite covered the client-side `_meta` logLevel stamp but not that a log actually flows back — the legacy suite already has that (`should receive server-initiated notifications via SSE/streamable-http`), so the modern side was the gap. Add two live e2e tests against a modern (`createMcpHandler`) server serving the `send_notification` tool with `logging: true`: - opted-in (server setting = a level): calling the tool delivers a `notifications/message` over the request's SSE stream, with the right level/logger/data. - not opted in (setting "off"): the tool still returns a result, but the server gates the log — no `notifications/message` is received. This exercises the full path end to end: the client seeding + `_meta` opt-in, and the server's `send_notification` → `ctx.mcpReq.log` era-correct gating. Verified with the full `npm run ci` (green across all clients). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * web: hide the modern "Log Level per Request" setting on legacy connections (#1629) The per-request log level is meaningless on a legacy server (which uses logging/setLevel), so hide it in Server Settings when the era is legacy: - protocolEra pinned to "legacy" → hidden (no connection needed), or - protocolEra "auto" that negotiated legacy at connect time → hidden. It stays visible for a pinned "modern" server, and for an "auto" server that is either not yet connected (era unknown) or resolved to modern. Threads the negotiated era down: App passes `negotiatedEra` to ServerSettingsModal only when the settings target is the live-connected server (`connectionStatus === "connected" && settingsModalTargetId === activeServerId`); the modal forwards it to ServerSettingsForm, which gates the control on `configuredEra === "modern" || (configuredEra === "auto" && negotiatedEra !== "legacy")`. Tests: visibility matrix on the form (legacy hidden; auto+legacy hidden; auto+unconnected and auto+modern and modern shown) and a modal hide case; the existing modern-log tests now base on a modern-era settings object. Verified in the browser (legacy hides, switching to modern reveals it reactively) and with the full `npm run ci` (green across all clients). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-19 17:08:22 -04:00
Cliff Hall	b3b04c6939	Thread cacheMode through managed list refresh via SDK cache-aware verbs (#1721) (#1723) ... * Thread cacheMode through managed list refresh via SDK cache-aware verbs (#1721) Manual/auto refresh of tools/resources/prompts now forces a cache-bypassing round trip on modern servers, delivering the raw-request cacheMode control deferred from #1628. The InspectorClient list/read path uses the SDK's single-page `client.request` verbs (for pagination debugging), which never consult the SDK response cache — so `cacheMode` was inert there. This adds a separate aggregate path that routes through the SDK's cache-aware high-level verbs, where `cacheMode` is honored: - InspectorClient: new `listAllTools/Prompts/Resources/ResourceTemplates( { cacheMode, metadata })` that call `client.listTools()` etc. (all-page aggregate + SEP-2243 x-mcp-header exclusion + cache consult/write). A shared `getCacheableRequestOptions` / `aggregateListParams` back all four. - InspectorClientProtocol: the four `listAll*` methods added. - ManagedListState: delegates page walking to the SDK verb (config `fetchPage` loop → `fetchAll`); `refresh(metadata?, cacheMode?)` threads cacheMode. A `list_changed` auto-refresh bypasses the cache (`cacheMode: "refresh"`). - Managed hooks: user-initiated `refresh()` passes `cacheMode: "refresh"`, so the existing per-tab Refresh buttons become genuinely cache-bypassing — no App wiring change. Connect-time load stays default (`'use'`). The single-page `listTools(cursor)` primitive is unchanged; the paged hooks still surface per-page `nextCursor`. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * Single-page list pagination: sidebar toggle + server setting (#1721) Delivers the deferred single-page fetch UI: Tools/Resources/Prompts can be loaded one page at a time instead of auto-aggregating every page — a defensive option for servers with very large lists. - New `singlePageLists` server setting (persisted, server-wide), plumbed through ServerConfig/InspectorServerSettings, serverList convert/default, the node + remote settings paths, and the Server Settings form ("Fetch Lists One Page at a Time"). - New `ListPaginationControls` element (a "Single page" switch + "Load next page" button + "N pages loaded" status) rendered in each list sidebar; the toggle edits the setting live (optimistic) and persists it. - The paged state stores (PagedTools/Prompts/ResourcesState) now own loading: they auto-load page 1 on connect in single-page mode and track nextCursor + page count as observable state; the managed stores skip their all-page walk in single-page mode (so nothing is over-fetched) while still lighting the list-changed indicator. `usePaginatedList` selects the display source and pagination controls per mode — pure, no data-loading in React effects. - Refresh is mode-aware: reload page 1 in single-page mode, refetch the whole aggregate otherwise. Sidebars widened 340→360 so the controls don't crowd list entries. - Test server: `test-servers/configs/pagination-http.json` (12 tools/resources/ prompts, page size 4 → 3 pages each) exercises the flow end to end. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * Add pr-screenshots proving single-page pagination (#1721) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * Address @claude review: template load, double-click guard, auth recovery (#1721) - Fix: resource templates never loaded on connect in single-page mode. The connect-time all-page skip now applies only to lists with a paged counterpart (new `deferInSinglePage` flag on ManagedListConfig; false for resource templates), so templates still aggregate on connect regardless of the setting. - Guard `PagedTools/Prompts/ResourcesState.loadPage` against a concurrent load (fast double-click on "Load next page") so the same page can't be appended twice; the in-flight load owns the cursor and the second call is a no-op. - Thread ambient auth recovery through the single-page refresh, Load-next-page, and mode-switch loads (they previously bypassed it) — usePaginatedList's onRefresh/onLoadMore now return their promises so App can wrap them. - Drop the misleading "aggregated list" wording from the paged-store docs. Tests added for each; npm run coverage green. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * Review round 2: revert optimistic toggle when persisting the setting fails (#1721) Non-blocking review follow-up. If updateServerSettings rejects, the optimistic `singlePageListsOverride` is now cleared and the live client setting rolled back (the effect only clears the override when the persisted value changes, which doesn't happen on failure), so the UI and client reflect what's on disk rather than the failed edit. Added an App test for the revert path. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * Fix single-page Refresh not clearing list-changed indicator + UI polish (#1721) Round-3 review fix (blocking): in single-page mode a `list_changed` lights the list-changed indicator, but the single-page Refresh reloads page 1 of the paged store and bypassed `useManaged*.refresh()`, so the indicator was never cleared. Expose `clearListChanged` from useManagedTools/Prompts/Resources and call it in the single-page branch of onRefreshTools/Prompts/Resources. Tests added. UI polish (user feedback on the sidebar control): - Rename the switch label "Single page" → "Paginated". - Put "Load next page" on the same row as the switch, space-between (switch fully left, button fully right). - Move the "N pages loaded" status to its own line below the row, centered. - Hide the whole control once the list is known to be a single page (single-page mode, page 1 loaded, no nextCursor) — nothing to paginate. pr-screenshots refreshed to the new layout. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * Rename the singlePageLists setting to paginatedLists for clarity (#1721) `singlePageLists` read backwards from the "Paginated" switch label (it sounds like paginated=false). Rename the server setting to `paginatedLists` with the same value meaning — true = paginated/one-page-at-a-time, false = aggregate-all — and follow through on the related identifiers for consistency: - Setting field `paginatedLists` across ServerConfig / InspectorServerSettings, serverList convert/default, node + remote settings paths, and the form/modal callbacks (onPaginatedListsChange). - `usePaginatedList` / `ListPaginationControls` prop `singlePage` → `paginated`, `onSinglePageChange` → `onPaginatedChange`. - `ManagedListConfig.deferInSinglePage` → `deferWhenPaginated`. - App state (paginatedListsOverride, onTogglePaginatedLists, …) and all tests/fixtures + mode-referring comments. "single-page" wording that denotes the SDK's one-page `client.request` verb (inspectorClient) is intentionally left as-is. Purely a rename — no behavior change; unit + integration + cli + tui + coverage all green. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * Format App.test.tsx (prettier) after the paginatedLists rename (#1721) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-19 01:06:31 -04:00
Cliff Hall	e026051b47	Era-aware connection model & capability gating (#1626) (#1697) ... * Era-aware connection model & capability gating (#1626) Make protocol era (SEP §7.8) a first-class, per-server connection dimension, orthogonal to the transport type, and drive the SDK Client's versionNegotiation from it. Config - Add `ServerProtocolEra` ("legacy" | "auto" | "modern"), `DEFAULT_PROTOCOL_ERA` (legacy, per the SDK's don't-auto-probe guidance), `MODERN_PROTOCOL_VERSION`, and `eraToVersionNegotiation()` in core/mcp/types.ts. - Persist `protocolEra` on `StoredMCPServer` / `InspectorServerSettings`, round-tripping through serverList.ts (omitting the default on disk) and the `/api/servers` read + write validators. InspectorClient / state - Thread `versionNegotiation` through `InspectorClientOptions` (default legacy), replacing the hardcoded pin. After connect, read `getProtocolEra()` / `getDiscoverResult()` / `getNegotiatedProtocolVersion()`, expose them via getters, dispatch `protocolEraChange` / `discoverResultChange`, and clear on disconnect. `useInspectorClient` surfaces `protocolEra` / `discoverResult`. Web UI - ServerSettingsForm: a Protocol Era selector bound to the per-server setting. - ConnectionInfoContent: era-aware — Era badge, Session ("Sessionless" on modern), and a Discovery section (supported versions + extensions) from the `server/discover` result. CLI/TUI also map the mcp.json `protocolEra` into versionNegotiation so the field applies uniformly. Version negotiation stays fully client-side; modern connect works over the existing remote tunnel and disconnect is sessionless (the SDK skips DELETE when there is no session id). Mcp-* header fidelity is left to the History/Network card. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * Address @claude review: era docstrings, era validation, stdio session label, AGENTS.md rule - Fix stale docstrings/comments in inspectorClient.ts, useInspectorClient.ts, inspectorClientEventTarget.ts, and ConnectionInfoContent.tsx: getProtocolEra() returns "legacy" (not undefined) on a plain legacy connect; it is undefined only when not connected. - Validate protocolEra at the storedFieldsToInspectorSettings read boundary (new shared isProtocolEra guard) so a hand-edited mcp.json read directly by the CLI/TUI drops an unknown era instead of passing it to eraToVersionNegotiation. Keeps the exhaustive switch intact. - ConnectionInfoContent: the Session row is HTTP-only — show "N/A (stdio)" for stdio instead of the misleading "Session-based". - AGENTS.md: add a rule against unjustified `as unknown as` double casts, and drop the one such cast introduced in the serverList test. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 * Address re-review: dedupe isProtocolEra guard in remote server Replace the local isProtocolEra in remote/node/server.ts with the shared guard now exported from serverList.ts (already imported here), so the era literal set lives in one place and can't drift. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5 --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-16 23:39:29 -04:00
cliffhall	8a7a38f800	Fix MCP App load regression: translate ext-apps v1 setNotificationHandler API to v2 ... An MCP App (e.g. NightRider's drum-pad) failed to load with "'[object Object]' is not a spec notification method; pass schemas as the second argument to setNotificationHandler()". `@modelcontextprotocol/ext-apps` still peers on SDK v1 and, during the AppBridge connect handshake, subscribes to tools/resources/prompts list_changed with the v1 schema-first API `setNotificationHandler(NotificationSchema, handler)`. SDK v2 requires a method STRING and throws on a schema object — the migration's compat shim covered the types but not this runtime API difference. Fix in the app-renderer client Proxy (`getAppRendererClient`): translate a schema-first `setNotificationHandler` call to v2's method-string form by reading the method literal off the schema's `shape.method.value`; native string-first calls pass through, and an unrecognized first arg falls through to the SDK's clear error rather than being swallowed. Only `setNotificationHandler` needed this — AppBridge's `request(req, schema, …)` calls work as-is (the v1 SDK schemas expose `~standard`, so v2 accepts them). Remove when ext-apps#702 ships a v2 peer. Added an integration test covering the schema-first translation end-to-end (handler fires on a real tools/list_changed), plus the string-passthrough and unrecognized-schema branches. Coverage ≥90 holds; typecheck + App tests green. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-15 12:44:20 -04:00
cliffhall	19a6df7ad3	SDK v2 migration (behavior-neutral, versionNegotiation: 'legacy') (#1624) ... Migrate all three clients + core + test-servers from @modelcontextprotocol/sdk@1.29 to SDK v2 (2.0.0-beta.4), keeping behavior identical against legacy (2025-11-25) servers. Foundation for the SDK V2 + New Spec stack. Packages - Split @modelcontextprotocol/sdk -> @modelcontextprotocol/client (Client, transports, OAuth, error taxonomy, protocol types) + core (Zod schemas); stdio -> client/stdio, Ajv -> client/validators/ajv. Test-servers -> @modelcontextprotocol/server (+ server-legacy/sse). Deterministic import rewrite (schemas -> core, everything else -> client); the codemod misroutes client code to /server so it was not used for path routing. - new Client(..., { versionNegotiation: { mode: 'legacy' } }) — nothing 2026-era goes on the wire. Runtime (core/mcp/inspectorClient.ts) - Error taxonomy: McpError -> ProtocolError, ErrorCode -> ProtocolErrorCode; RequestTimeout -> SdkError/SdkErrorCode. - Handlers take method strings; task methods (tasks/*, notifications/tasks/*) use the 3-arg custom form (excluded from v2's spec-method set). - Legacy tasks kept working (per project requirement): requestor tasks re- implemented over raw client.request + explicit deprecated schemas; experimental.tasks.callToolStream replaced by a local pollTaskToolCall async generator (with _progressHandlers revival so task-execution progress still reaches the caller). Receiver (server-initiated) tasks: the v2 Client validates a sampling/elicit handler result and rejects the legacy { task } response with -32602 — installReceiverTaskResponseBypass swaps the wrapped entry in the private _requestHandlers map for the task-augmented branch so { task } reaches the wire (fixes server-initiated sampling-via-task). - Pagination: list verbs use raw client.request({method}, List*ResultSchema) for single-page behavior (v2's client.listX() auto-aggregate all pages). Web / auth / apps - Rename local JsonSchemaType -> InspectorFormSchema + add toFormSchema() structural narrowing at the SDK-schema -> form boundary. - ext-apps still peers on SDK v1: one compat boundary with documented casts (remove when ext-apps#702 lands a v2 peer). - OAuth provider saveClientInformation reconciled with v2's OAuthClientInformationContext; InvalidClientMetadataError re-homed as a local class (server-legacy/auth is not browser-safe). - vitest.shared.mts dedupes @modelcontextprotocol/client+core to one copy so vi.mock intercepts the SUT's import. Accepted v2 behavior changes (per #1624 / spec §9.2): unknown-tool call rejects with -32602; cancellation is stream-abort (no guaranteed notifications/cancelled frame); output-schema strictness restored via the Inspector's own Ajv check on the default path. npm run ci green: validate (web 3087 / cli 212 / tui 274 / launcher 5), coverage (>=90 per-file, all clients), smoke (launcher/cli/web), storybook (410). Closes #1624 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-15 10:44:24 -04:00
cliffhall	32a3610737	Merge remote-tracking branch 'origin/v2/main' into 1579-wave-1 ... # Conflicts: # clients/cli/package-lock.json # clients/cli/package.json # clients/web/src/App.tsx # clients/web/src/lib/environmentFactory.test.ts # clients/web/src/lib/environmentFactory.ts # clients/web/src/lib/remoteOAuthStorage.test.ts # clients/web/src/lib/remoteOAuthStorage.ts # clients/web/src/test/core/auth/connection-state.test.ts # clients/web/src/test/core/auth/providers.test.ts # clients/web/src/test/core/auth/storage-browser.test.ts # clients/web/src/test/core/auth/storage-remote.test.ts # clients/web/src/test/core/mcp/oauthManager.test.ts # clients/web/src/test/integration/auth/node/storage.test.ts # clients/web/src/test/integration/storage/adapters.test.ts # clients/web/src/utils/clearServerOAuthState.test.ts # clients/web/src/utils/clearServerOAuthState.ts # core/auth/connection-state.ts # core/auth/oauth-storage.ts # core/auth/providers.ts # core/auth/storage.ts # core/auth/store.ts # core/storage/adapters/remote-storage.ts	2026-07-09 09:41:08 -04:00
cliffhall	cfbdfdf2bd	Merge remote-tracking branch 'origin/v2/main' into 1579-wave-1 ... # Conflicts: # AGENTS.md # clients/cli/README.md # clients/web/src/App.test.tsx # clients/web/src/App.tsx # clients/web/src/test/core/auth/utils.test.ts # core/mcp/node/transport.ts # test-servers/src/composable-test-server.ts	2026-07-03 17:39:07 -04:00
Bob Dickinson	a24091484c	fix(auth): await OAuth storage load() to fix remote OAuth races	2026-07-03 14:18:00 -07:00
Bob Dickinson	562bdd00f1	All of the mid-session and step-up auth work	2026-07-02 16:35:31 -07:00
cliffhall	4bad110942	test(web): eliminate timeout flakiness in the web test suite (closes #1596) ... Root-cause and deterministically fix the recurring 5s-timeout flakes that surfaced under v8-instrumented, concurrent load. All fixes are test-only; no product source changed. No test was skipped or disabled, and the per-file coverage gate (>=90 on all four dims) still holds. Common root cause across the modal/form/screen suites: userEvent.setup() schedules a real setTimeout between keystrokes. Under CPU contention those yields balloon, so multi-field interactions blew past the 5s per-test timeout. Fix: userEvent.setup({ delay: null }) removes the per-keystroke real-timer dependence (typing dispatches synchronously), making the suites load-independent. Per file: - ServerConfigModal / ServerImportConfigModal / ServerImportJsonModal / ResourcesScreen / InspectorView: userEvent.setup({ delay: null }). - PromptArgumentsForm (completions): replaced wall-clock sleeps sized just past the 300ms completion debounce (await setTimeout(400)) — which race the debounce timer under load — with awaited conditions (findBy/waitFor) on the real rendered outcome, plus delay:null. Negative assertions now lean on the component's synchronous state/timer teardown instead of a timed window, so they are deterministic regardless of machine speed. - inspectorClient integration "tracks stderr logs": the child's stderr is piped out-of-band from the tool's JSON-RPC response, so reading the log synchronously after callTool raced the stderr chunk. Wrapped the assertion in vi.waitFor so it polls until the line lands. Verified: web validate + 6x test:coverage (4 sequential + 2 concurrent under load) + standalone test:integration, all green (229 files / 3237 tests per coverage run; 799 integration tests), zero intermittent failures. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01S3fTN8H3R8YV4yUGvZjYnX	2026-07-01 21:11:11 -04:00
Bob Dickinson	13dafe457b	Guided-auth removal, OAuth smoke testing, and TUI/CLI support (#1531) ... * refactor(web): update OAuth E2E tests to use new authentication method - Replaced `runGuidedAuth` with `authenticate` in the InspectorClient OAuth end-to-end tests to support quick authentication with CIMD pre-registration. - Updated comments to reflect the change in authentication method and its support for HTTP test metadata URLs. - Ensured that the tests continue to validate the expected authorization URL after the change. Additionally, imported `ensureCimdClientRegistration` in the OAuthManager to streamline client registration during the authentication process. * Guided-auth removal, OAuth smoke testing, and TUI/CLI support * Added CLI OAuth runner flags tests * Added clarifying comment on port 0 binding * README notes on TUI oauth callback port, prettier fixes for CLI tests * CIMD validation uses inline like EMA IdP * Fixed TUI test failures from merge * test: bring 5 OAuth files over the 90 per-file coverage gate The guided-auth-removal work added/touched several files that fell below the repo's >=90 per-file coverage gate (npm run coverage). CI doesn't run the gate (it's local-only), so validate + test:integration passed while these gaps slipped through. Bring them all back over the line: - core/auth/node/runner-oauth-callback.ts: test bad-URL/non-http-scheme throws, no-port->80 default, and IPv6 hostname bracketing. Annotate the provably-dead guards (empty host, out-of-range port, String(err) fallback, pathname "/" fallback) -- all rejected by new URL() upstream -- with justified v8 ignore comments instead of lowering the gate. - core/auth/cimd.ts: test the AS-doesn't-support-CIMD early return. - core/react/useClientSettingsDraft.ts: test the functional-updater path, the stale-onChange-after-close prev===null guard, and the timer-fires-after-close value===null guard. - ClientSettingsForm.tsx: test the CIMD checkbox toggle and CIMD URL clear. - clients/tui/src/utils/oauthDisplay.ts: test static/dcr registration kinds, the expired IdP session, and the configured-scope/empty-scope branches of formatScopes. npm run coverage now passes (exit 0) across web, cli, tui, launcher. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: cliffhall <cliff@futurescale.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-06-29 15:35:51 -04:00
Bob Dickinson	5de4e911e3	feat(web,core): enterprise-managed authorization (EMA) (#1509) (#1512) ... Adds enterprise-managed authorization (EMA): browser-side IdP login + RFC 8693 token-exchange and JWT-bearer resource-token mint, per-server EMA opt-in in Server Settings, IdP config in Client Settings, and IdP clientSecret keychain storage. Includes the v2/main merge and review follow-ups (400 on malformed client config; normalizeIdpIssuer import consolidation).	2026-06-24 22:15:12 -04:00
Cliff Hall	5ee347571f	feat(web,core): persist & edit STDIO cwd/env in Server Settings (#1515) (#1518) ... * feat(web,core): edit STDIO cwd/env in Server Settings (#1515) Persist and edit a STDIO server's working directory and environment variables through the Server Settings modal. - Add `env` (pair-array, mirrors headers) and `cwd?` to `InspectorServerSettings`. - `storedFieldsToInspectorSettings` lifts the stdio config's `env`/`cwd` into settings so the modal can show them; they stay on `config` for the transport. Add `envRecordToPairs` / `envPairsToRecord` helpers. - `env`/`cwd` remain config-owned on disk (the standard mcp.json location): the `/api/servers` PUT route writes an edited settings.env/cwd back onto the stored stdio config on a settings-only patch (supporting clear), and a config-providing PUT keeps config authoritative. `validateSettings` accepts `env`/`cwd`. - Server Settings modal: Working Directory field in Options and a dedicated collapsible Environment Variables section — both STDIO-only. - Registry server.json import sets no `cwd` (the schema has no working-directory concept); documented in serverJson.ts. Tests: serverList converter round-trip, route write-through (set/clear, config-precedence, non-stdio no-leak, validation), import preservation (client-config cwd, registry cwd skip), and form/modal component coverage. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(web): preserve config env/cwd when a settings PUT omits them (#1515 review) Address Claude's review of #1518. - `/api/servers` PUT now distinguishes an *absent* `env`/`cwd` (preserve the stored value) from an explicit empty list / blank string (clear). Previously `validateSettings` coerced a missing `env` to `[]`, so an env-unaware caller patching only a timeout would silently wipe the server's env/cwd. The integrated web client always resends the full GET-rehydrated env, so it was unaffected — this protects the raw HTTP contract. - `validateSettings` reports `envProvided`/`cwdProvided`; the write-through only touches `config.env`/`cwd` for fields the caller actually sent. - On the omit path, rehydrate the untouched env's keychain values onto the rebuilt entry so the secret reconcile re-persists them rather than sweeping the blanked-on-disk keys out of the keychain. - Minor: guard the stdio cast in `mcpConfigToServerEntries` on the transport type instead of casting unconditionally. - Test: settings-only PUT omitting env/cwd preserves config keys + cwd + the keychain secret. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-06-24 19:15:43 -04:00
Cliff Hall	dc31c28102	fix(core): make "error" the canonical terminal status on a mid-session crash (#1490) (#1493) ... * fix(core): make "error" the canonical terminal status on a mid-session crash (#1490) On a real mid-session transport crash many SDK transports fire BOTH `onclose` and `onerror` in a transport-dependent order. The `onclose` guard (`if status !== "disconnected"`) let a trailing `onclose` downgrade a crash's `"error"` status back to `"disconnected"`, so the persistent terminal status differed by ordering: - onerror-then-onclose → "disconnected" - onclose-then-onerror → "error" `onclose` now treats `"error"` as terminal (only `"connected"`/ `"connecting"` → `"disconnected"`), so `"error"` is the canonical resting status in both orderings. It still emits the `disconnect` event exactly once so session-teardown consumers fire identically either way. Because the state managers gated list/message/subscription/task teardown on `getStatus() === "disconnected"`, holding the status at `"error"` would have left them stale in the onerror-first ordering — swapping one ordering-dependence for another. Introduce `isTerminalStatus()` (disconnected || error) and key every teardown guard off it so they clear identically regardless of crash ordering. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(core): single disconnect event when explicitly disconnecting from "error" (#1490 re-review) Addresses the non-blocking observation from the PR review: explicitly calling disconnect() from a held-"error" status could fire the `disconnect` event twice. close() can trigger the transport's onclose synchronously; onclose (holding status at "error") fired `disconnect`, then disconnect()'s own guard saw status still != "disconnected" and fired it again — a timing-dependence on whether the SDK closes sync. Add a `disconnecting` ownership flag: onclose defers its status set + `disconnect` event to disconnect() while it is set, so the explicit-disconnect path emits exactly one `disconnect` regardless of sync/async close. Crash-path onclose (flag unset) is unchanged. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * test(core): deterministically pin the sync-close disconnect double-fire (#1490 re-review) The real-stdio test asserts the single-disconnect invariant but may not exercise the sync-close branch (the one that double-fired) if the SDK stdio transport closes asynchronously. Add a test that overrides the SDK client's close() to invoke the transport's onclose synchronously, forcing the exact timing the `disconnecting` flag guards. Verified it asserts 2 (fails) without the flag and 1 (passes) with it. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-06-15 10:15:29 -04:00
Cliff Hall	b4716a1405	feat(core): surface mid-session transport failures via lastError (#1489) ... Closes #1323. Adds an `error`-event-driven `lastError` on useInspectorClient and a mid-session failure toast in App.tsx; the transport onerror is scoped to non-handshake (status !== connecting) so handshake failures stay on the connect() rejection path. Follow-up: #1490.	2026-06-15 08:46:59 -04:00
Cliff Hall	53dc4549be	fix(core): trace dropped fetch body update when entry rotates out (#1390) (#1480) ... * fix(core): trace dropped fetch body update when entry rotates out (#1390) FetchRequestLogState.onFetchRequestBodyUpdate silently dropped a deferred response-body update when the matching entry had already been evicted by log rotation (>= maxFetchRequests newer requests appended between the append and the fire-and-forget body read). The user saw an empty Response Body with no way to distinguish "body never came back" from "body dropped after eviction". Plumb an optional pino logger into FetchRequestLogState (defaulting to the silent logger) and emit a logger.debug at the early-return site so the drop is traceable. The web client wires the existing remote logger from createWebEnvironment into the state manager. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(core): emit dropped-body trace at warn level so it is observable (#1390) Review feedback: the web remote logger (createRemoteLogger) defaults to level 'info' with no override at the environmentFactory call site, so a logger.debug at the drop site is filtered out by both the pino level gate and transmit.level — the diagnostic never reaches /api/log and the drop stays invisible in the shipped web app, defeating the issue's goal. Promote the trace to logger.warn (>= info, so it transmits by default). A response body lost to log rotation is unexpected data loss, not routine diagnostic noise, so warn is the appropriate severity. Update the comment to explain the level choice and switch the test stub from debug to warn, extracting a small typed makeWarnLogger helper per review. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(web): configurable Network Log Size per server + body-dropped toast (#1390) Turn the otherwise-invisible "response body dropped after rotation" case into an actionable signal: 1. Per-server `maxFetchRequests` setting (default 1000, 0 = unlimited), editable as "Network Log Size" in the first (Options) accordion section of the Server Settings form. Plumbed through the full settings pipeline: InspectorServerSettings + StoredMCPServer types, the serverList disk converters (read defaults to 1000, write omits the default to keep diffs minimal — 0 is a meaningful non-default), and the node server's GET-sanitize + write-side validateSettings (non-negative number). Applied live to the active connection on settings-modal close via a new FetchRequestLogState.setMaxFetchRequests (shrinking trims immediately), and at connect time for the next session. 2. FetchRequestLogState now emits a `fetchRequestBodyDropped` event, but only for genuine rotation drops (log at capacity with a finite cap) — not for benign post-clear stragglers or unlimited logs, which previously also hit the warn path. App shows a deduped, persistent toast (stable per-server id, no auto-close) explaining what happened and the likely causes, with a link that opens this server's settings on the Options section (modal remounts via key) so the user can raise the limit if it's just a high-traffic server. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(web): don't treat a cleared Network Log Size field as unlimited (#1390) Review feedback: clearing the NumberInput emits "", and `parseInt("") || 0` coerced that to 0 (unlimited). Since the value applies live on modal close, a user who cleared the field to retype and then closed the modal would silently switch the log to unlimited. Fall back to the current value on an empty/NaN parse and reserve 0 for an explicit numeric entry. Adds a cleared-input test. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-06-14 17:35:52 -04:00
Cliff Hall	3220d6239a	fix(web): send a cancellation notification when cancelling an ordinary tool call (#1458) (#1463) ... * fix(web): send a cancellation notification when cancelling an ordinary tool call (#1458) The Tool detail panel's Cancel button was a silent no-op for ordinary (non-task) tool calls: App.onCancelToolCall found no captured taskId and did nothing, leaving the button looking actionable while the call ran on. Wire request cancellation through the MCP cancellation flow. callTool now tracks the in-flight call with an AbortController whose signal is threaded into the SDK request options; aborting it makes the SDK send a notifications/cancelled for that request and reject the pending call. cancelToolCall() performs the abort. callTool surfaces the cancellation as a ToolCallCancelledError (without recording a failed call), which App clears as a clean cancel plus a short confirmation toast. Task-augmented calls are unchanged: they have a server-side task and are still cancelled via cancelRequestorTask (#1457). Closes #1458 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(web): address review — distinguish disconnect from cancel; tidy callTool params Review follow-ups on #1463: - Disconnect no longer surfaces as "Tool call cancelled". The catch in callToolWithRetries now throws ToolCallCancelledError only when the abort reason matches the deliberate user-cancel reason (TOOL_CALL_CANCELLED_REASON); a disconnect aborts the same controller with a different reason and falls through to the normal error path, so no misleading toast. - Note the activeToolCallAbortController slot is shared (last-writer-wins) across all callTool callers. - Fold callToolWithRetries/attemptToolCall's long positional parameter list into a single ToolCallRequest object; move the URL-elicitation loop state into callToolWithRetries. Adds an integration test that a disconnect mid-call rejects as an ordinary error (not ToolCallCancelledError), using a slow send_progress call over HTTP. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-06-13 22:23:02 -04:00
Cliff Hall	6bb49e55ba	fix(web): cancel the underlying task when cancelling a task tool call (#1455) (#1457) ... * fix(web): cancel the underlying task when cancelling a task tool call (#1455) The ToolDetailPanel Cancel button fired onCancel, but App never supplied an onCancelToolCall handler, so clicking Cancel did nothing — a tool running as a task kept running on the server. App now captures the in-flight task-augmented call's taskId from the `toolCallTaskUpdated` event callToolStream dispatches, and the new onCancelToolCall handler cancels that task via the existing tasks API (reusing onCancelTask). The id is reset at the start of every call, so an ordinary (non-task) call's Cancel stays a no-op (no task to cancel, no request-abort channel yet) — preserving prior behavior. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(web): show a toast when a task is cancelled Cancellation goes out as a `taskCancelled` event (not a status notification), so a running task's open 'Task <status>' toast would otherwise hang with no confirmation. Subscribe to `taskCancelled` and replace that toast (or show a fresh one) with a short auto-closing 'Task cancelled' toast, pruning the dead progress entry. Covers both cancel paths — the Tasks screen and the Tool detail panel. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(core): report a cancelled task as cancelled, not failed When a task is cancelled, the in-flight callToolStream rejects with a generic -32603 error, whose error path reported the task as "failed" — so the Tasks screen flashed FAILED until a manual refresh fetched the server's true "cancelled" state. InspectorClient now records cancelled task ids (set in cancelRequestorTask before the await, since cancelling unblocks the stream) and the stream's error path labels the terminal task "cancelled" with a matching status message. Cleared on disconnect. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * refactor: address review nits on task-cancel wiring - Prune the consumed taskId from cancelledTaskIds in the stream error path so the set can't grow across a long session (disconnect-clear remains the backstop for cancels whose task completed before the cancel landed). - Clear activeToolCallTaskIdRef before the cancel resolves so a rapid second Cancel click doesn't re-cancel the terminating task (which would surface a spurious 'Failed to cancel task' toast). Add a test. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-06-13 18:00:48 -04:00
Cliff Hall	721dae08c6	feat: apply auto-refresh setting live + diff-aware list-changed indicator (#1444) (#1445) ... * feat: apply auto-refresh setting live + diff-aware list-changed indicator (#1444) Two follow-ups to #1402. 1. Live setting application — InspectorClient gains setServerSettings() (on the protocol + fake); App pushes edited settings onto the live client when the Server Settings modal closes for the connected server, so toggling auto-refresh-on-list-changed takes effect without a reconnect (the managed state reads getServerSettings() at notification time). 2. Diff-aware indicator — in flag-only mode, on */list_changed the managed Tools/Prompts/Resources state now fetches the new list and compares it to the current one, lighting the indicator only when it actually differs. The displayed list stays put until the user pulls via Refresh (pull-on-demand preserved). Servers like everything re-send identical lists on list_changed; this suppresses the false-positive indicator. Refactors each manager's refresh into fetch + apply helpers shared with the peek path. Trade-off: item 2 reintroduces a fetch on */list_changed (to compare), but unlike the pre-#1402 behavior it never replaces the displayed list under the user — only the indicator reflects a real change. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * refactor(core): make the diff-aware peek clear the indicator on revert (#1444 review) peekForChange now sets the list-changed flag to whether the server list differs from the displayed one (setListChanged(differs)) instead of only ever setting it true. So a later notification that reverts the server back to the displayed list clears the indicator — nothing left to pull. Adds a clear-on-revert regression test. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * refactor(core): consolidate managed-list state managers onto a shared base (#1444) The four per-primitive list managers (tools, prompts, resources, resource templates) were near-identical copies of the same fetch/apply/peek/connect/ disconnect/list_changed machinery, a standing drift risk. Extract a generic ManagedListState<T, M> base parameterized by a config (change event, list_changed notification, capability key, page fetcher, and whether the list drives an indicator). Each manager is now a ~40-line subclass: its config plus a typed getter alias. Public contracts are unchanged — same per-manager *Change / listChangedChange events and get*/refresh/getListChanged/clearListChanged methods — so the hooks and all existing tests pass untouched. Templates opt out of the indicator (supportsIndicator: false): a list_changed auto-refreshes only when the server opts in, else does nothing. Net -245 lines. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * perf(core): coalesce bursts of list_changed peeks (#1444 review) A rapid burst of */list_changed notifications fired one concurrent paginated fetch per notification. Guard peekForChange with an in-flight flag: while a peek is fetching, a new notification queues a single re-run instead of launching a concurrent fetch, and the running peek loops once more on completion to capture the latest server state. Collapses a burst into sequential peeks. Adds a coalescing test. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(core): debounce list_changed bursts into a single fetch (#1444) The in-flight guard only coalesced fetches that actually overlapped — when a server (e.g. everything) emits a burst of */list_changed spaced just far enough apart that each list call returns before the next notification, we still fired one list call per notification. Add a real debounce in the ManagedListState base: a notification (re)starts a timer and the refresh/peek runs once the burst settles, collapsing N notifications into one fetch. The in-flight coalescing guard stays as a second line of defense for a post-debounce notification landing during a slow paginated fetch. The delay is injectable (DEFAULT_LIST_CHANGED_DEBOUNCE_MS = 250; subclasses take an optional constructor arg) so tests run with 0 for fast, determinist ic behavior. Timer is cleared on disconnect, unsubscribe, and destroy. Adds debounce + in-flight regression tests. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * refactor(core): extend the in-flight guard to the auto-refresh path (#1444 review) The overlap guard was only on the peek path; a burst-during-a-slow auto- refresh could let a slower older fetch clobber a newer list (applyItems is last-write-wins). Lift the guard from peekForChange up to runListChanged so it wraps both the auto-refresh and peek actions: while one is fetching, a new post-debounce notification queues a single re-run. Adds an auto-refresh coalescing test alongside the peek one. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(core): light the indicator without fetching when auto-refresh is off (#1444) Per maintainer feedback: with auto-refresh off, a list_changed should not issue any list call. The diff-aware peek (fetch-to-compare) meant one tools/list went out per settled burst even with auto-refresh off, which the maintainer doesn't want. Revert flag-only mode to lighting the indicator blindly on a (debounced) list_changed — no network — and the user pulls the new list via Refresh. Trade-off (accepted): the indicator lights even when the server re-sent an identical list (the everything-server case), since we no longer fetch to compare. Keeps the debounce (one indicator light per burst), the auto-refresh path + its in-flight guard, the live setServerSettings, and the ManagedListState consolidation. Removes peekForChange and its tests. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs(core): scrub stale peek docstrings after the blind-indicator revert (#1445 review) The diff-aware revert removed peekForChange's body but left its docstring, which orphaned onto destroy() and described fetch-to-compare/clear-on-revert behavior that no longer exists. Replace it with a one-line destroy() doc, and update clearListChanged's doc (peek -> blind light). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs(core): finish scrubbing stale peek references after the revert (#1445 review) Round-7's scrub missed four more peek/fetch-to-compare mentions in managedListState.ts (supportsIndicator, fetchAll, the debounce-default and onListChanged comments) plus six 'async peek' comments in the state tests. Update them all to the blind-indicator semantics. Doc/comment-only; no behavior change. grep -i peek over core/mcp/state now returns nothing. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-06-09 11:00:28 -04:00
Cliff Hall	9b558afee1	feat(web): display negotiated MCP protocol version on ServerCard (#1324) (#1442) ... * feat(web): show negotiated MCP protocol version on ServerCard (#1324) Add optional protocolVersion to ConnectionState and render it as a dimmed "MCP <version>" label at the end of the transport-badge/ description row on ServerCard, shown only when connected. The value is plumbed through useInspectorClient as part of #1324. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(core,web): surface negotiated protocolVersion end-to-end (#1324) Plumb the MCP protocol version negotiated during initialize from the InspectorClient through to the ServerCard so the active connected card displays "MCP <version>". The SDK Client doesn't expose protocolVersion; it stamps it onto the transport via setProtocolVersion. MessageTrackingTransport now captures it (and forwards to the base transport when present), so it works for stdio too — whose base transport has no setProtocolVersion. - core: MessageTrackingTransport captures + exposes protocolVersion; InspectorClient stores it, exposes getProtocolVersion(), dispatches a protocolVersionChange event, and clears it on disconnect. - core: useInspectorClient subscribes and returns protocolVersion. - web: App reads it from the hook (drops the hard-coded "2025-06-18") and InspectorView splices it onto the active server's connection. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * refactor(web): decouple initializeResult from protocolVersion (#1324 review) Gate the initializeResult memo on serverInfo only, not protocolVersion. initializeResult also drives the connected ViewHeader and Connection Info modal, so requiring the version risked hiding the whole connected surface in the edge case where it's absent. protocolVersion now flows through as the optional field it already is everywhere downstream ('' reads as unknown; the ServerCard label and modal value both tolerate it). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * refactor(web): drop redundant connected-check in InspectorView splice (#1324 review) initializeResult is derived from connectionStatus in App's memo, so a truthy initializeResult.protocolVersion already implies a live connection — the extra connectionStatus === 'connected' guard in the splice was redundant. ServerCard keeps its own status guard (dumb- component robustness). No behavior change. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-06-08 07:40:02 -04:00
Cliff Hall	49d6be6987	feat(web): add per-server Roots config + client roots capability (#1425) (#1431) ... * feat(web): add per-server Roots config + client roots capability (#1425) Add MCP roots support to the v2 web Inspector as a per-server config section in the server settings form (no separate Roots tab/screen). - Persistence: add `roots: Root[]` to `InspectorServerSettings` and `StoredMCPServer`; thread through serverList converters (lift on read, drop empty-uri rows / blank names on write, omit when empty) and the remote server's read-side drop guard + `validateSettings`. - Capability: `setupClientForServer` always advertises the roots capability (`roots: cleanRoots(savedSettings?.roots ?? [])`) so the server can issue `roots/list` and receive `roots/list_changed`. - Live apply: edits to a connected server's roots are pushed via `inspectorClient.setRoots(...)` once, on settings-dialog close — diffed against the client's current roots — rather than per keystroke, to avoid flooding the server with `roots/list_changed` / `roots/list`. - Form: new Roots accordion section with a required URI and an optional name per row (unlike v1, which was URI-only), + Add Root, and remove. Tests cover the persistence round-trip (serverList + servers-route), the form/modal section, and reuse the existing InspectorClient roots/`setRoots` integration coverage. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * refactor(roots): address review — dedup guards/cleanRoots, preserve extra fields, test live-apply Responding to the PR #1431 code review: - Dedup `isRootArray`: drop the copy inside `validateSettings` and reuse the module-scope guard the read path uses, so the two can't drift (core/mcp/remote/node/server.ts). - Dedup the clean-roots normalizer: hoist a single exported `cleanRoots` into core/mcp/serverList.ts and reuse it from both `inspectorSettingsToStoredFields` and App.tsx (drops the local copy). - Preserve non-form root fields (e.g. `_meta`): `cleanRoots` now spreads the carried-through fields and only normalizes uri/name, and the modal's `handleRootChange` spreads the existing root on edit — so a hand-edited root's `_meta` survives load + persist + edit instead of being silently dropped. - Test the App live-apply-on-close path: new App.test.tsx cases assert setRoots is called once with cleaned roots when changed on the active server, and not called when the cleaned roots already match what the client advertises. Adds a cleanRoots unit test (blank drop + _meta preserve). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-06-05 11:08:18 -04:00
Cliff Hall	ed6eda7492	feat(web): wire up MCP task support — Tasks screen, Run-as-task, status toasts (#1422) (#1424) ... Wires the ported Tasks screen + core task subsystem through App.tsx: sticky Clear Completed, cancel/refresh error toasts, core-side per-task progress correlation, Run-as-task on the Tools screen (gated on server support), per-server Task TTL, and live task-status toasts.	2026-06-04 08:49:47 -04:00
Cliff Hall	5fb2b503f4	feat(apps): render running MCP Apps end-to-end (#1370) (#1403) ... * feat(apps): render running MCP Apps end-to-end (#1370) Wire the v2 Apps screen so clicking Open App actually runs an MCP App. Previously App.tsx stubbed the three load-bearing pieces (sandbox URL, bridge factory, action handlers), so the iframe loaded about:blank and nothing crossed the bridge. - Add useSandboxUrl hook (core/react) to read sandboxUrl from GET /api/config and thread it down as sandboxPath; render an "MCP Apps unavailable" state instead of a blank iframe when absent. - Add createAppBridgeFactory: builds a host-side ext-apps AppBridge over the active client's SDK client, reads the tool's UI resource on the sandbox handshake and pushes it into the proxy, handles openLinks, and connects a PostMessageTransport. - AppRenderer: pass the tool to the bridge factory; buffer tool input/result and release them once the view signals `initialized` (input before result, latest-wins). - Move the AppRenderer handle ref to App.tsx so onOpenApp issues tools/call with the form values and delivers input + result to the running app via the bridge; onCloseApp tears the bridge/iframe down. - Stories: themed sandbox placeholder (visible in light/dark), an echo-bridge running story, and an unavailable story. New unit tests for the factory, the buffering, useSandboxUrl, and the empty state. The listChanged indicator badge wiring is tracked separately in #1402. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(apps): make MCP Apps render in dev — single bridge + result passthrough Two bugs kept apps (e.g. NightRider's MIDI Pattern Editor) stuck on an empty shell even though the host fetched the resource and called the tool correctly. 1. StrictMode double-mounted the bridge. The dev-only effect setup→cleanup→ setup built TWO host bridges/transports, double-loaded the sandbox, and raced the app's ui/initialize handshake (it timed out, so `initialized` never fired and buffered input/result were never flushed). AppRenderer now defers bridge disposal to a microtask and, when the effect re-runs synchronously with identical inputs, cancels the disposal and reuses the live bridge — so a single bridge survives the double-invoke. The pending input/result buffer is no longer cleared on a transient cleanup. 2. The SDK rejected the server's result. callTool() validates structuredContent against the tool's outputSchema and throws on a mismatch — but an MCP App is a passthrough: the host must forward whatever the server returned to the running view (which legacy hosts render fine). InspectorClient.callTool gains a `skipOutputValidation` option that routes through client.request() (no wrapper validation); onOpenApp uses it. Verified end-to-end against a real app via Playwright: the editor now renders its grid + pattern data identically to the v1 inspector. Tests: AppRenderer StrictMode single-bridge/no-dispose case; a get_temp_extra test-server fixture (returns an extra structuredContent property) with integration tests covering both the throwing default path and the skipOutputValidation passthrough. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(apps): warn (toast) when an app's result violates its outputSchema The skipOutputValidation passthrough keeps schema-mismatched apps rendering, but that silently hid a real spec problem from server developers. Now the host still renders, but surfaces the mismatch. - InspectorClient.callTool runs the SDK's own Ajv output validation on the skipOutputValidation path NON-fatally and returns the message as `outputValidationError` on the ToolCallInvocation (and the toolCallResultChange event); the result is still delivered. - App.tsx onOpenApp shows a yellow toast when outputValidationError is set, explaining strict MCP clients may not render the app. Tests: integration cases asserting outputValidationError is set for a schema-violating result (get_temp_extra) and unset for a valid one (get_temp). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(apps): summarize output-schema warning in toast + details modal The previous toast dumped the full (very long) validation error and overflowed. Now the toast shows a one-line summary with a "View validation details" link that opens a modal containing the full error in a read-only, scrollable textarea. - New OutputValidationModal group component (+ story + tests). - App.tsx: toast message is a summary + link; clicking it opens the modal populated with the tool name and the full outputValidationError. Verified via Playwright against the NightRider MIDI editor: toast is concise, link opens the modal with the scrollable details. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * refactor(apps): address PR review — theme stability, onError, validation polish Responds to the automated PR review on #1403. - Theme toggle no longer rebuilds a running app's bridge. createAppBridgeFactory reads the resolved theme from the DOM (Mantine's data-mantine-color-scheme) at bridge-build time instead of capturing isDark in the factory deps, so the factory identity is stable across theme flips. (#1) - Wire AppRenderer's onError through AppsScreen -> InspectorView -> App.tsx, surfacing bridge/runtime failures (no client, late rejection, failed tools/call) as a red toast instead of a silent blank iframe. (#2) - Extract validateToolOutput into a pure, unit-tested helper covering the matching / extra-property / missing-structuredContent / error / bad-schema cases (the no-structuredContent branch can't be produced via the SDK server, which rejects it before sending). (#4) - useSandboxUrl aborts its in-flight fetch on unmount (cancelled flag) to avoid setState after unmount. (#5) - Schema-mismatch toast: autoClose:false (details modal is one click away). (note) - AppRenderer: add a state-machine comment block documenting the bridge lifecycle refs. (#6) - inspectorClient: note why the `as CallToolResult` casts are safe. (note) Tests: AppsScreen onError surface; validateToolOutput unit suite; factory theme read from DOM. validate + test:storybook green. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs(apps): clarify theme attribute + useSandboxUrl fetch comments Re-review follow-ups (non-blocking): - createAppBridgeFactory: note that Mantine's data-mantine-color-scheme is only ever light/dark (auto is resolved before paint), so no `auto` branch is needed and matchMedia only covers an absent attribute. - useSandboxUrl: correct the doc to say it fetches on mount AND re-fetches on baseUrl/authToken change (idempotent), not "once on mount". Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-06-01 18:01:44 -04:00
Cliff Hall	440b194f64	fix(state): dispatch connect event after capabilities are populated (#1405) ... * fix(state): dispatch connect event after capabilities are populated The list-state managers (tools/prompts/resources/templates) refresh on the "connect" event and, since #1395, gate their list RPC on getCapabilities(). But connect() dispatched "connect" before fetchServerInfo() populated this.capabilities, and refresh() runs synchronously up to its first await — so the capability gate executed during the synchronous dispatch, read undefined capabilities (reset to undefined on every disconnect), tripped, and wiped every list to empty on connect. Tools, prompts, resources, and resource templates all vanished against any server. Move the "connect" dispatch to after the awaited fetchServerInfo() so capabilities are in place before any manager refreshes. The unit tests missed this because their FakeInspectorClient has capabilities populated from construction; added an integration regression test wiring a real ManagedToolsState on connect and asserting the tool list is non-empty (fails without the fix). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * test(integration): harden connect-race regression test per review Address the two non-blocking suggestions from the PR review: - Replace the 100ms settle sleep with waitForEvent() on the managers' change events, so the test is deterministic against slow CI instead of racing a fixed timeout. - Assert a second affected primitive (ManagedResourcesState alongside ManagedToolsState). The fix is a single shared line, but covering two distinct capabilities guards against a future change gating one primitive differently. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-06-01 11:10:07 -04:00
Cliff Hall	85a4b4c057	feat(servers): persist per-server settings to mcp.json (#1352) (#1353) ... * feat(servers): persist per-server settings to mcp.json (#1352) - Add optional `settings` node alongside the transport config in each `~/.mcp-inspector/mcp.json` entry. Round-trips through `mcpConfigToServerEntries` / `serverEntriesToMcpConfig` and the `/api/servers` POST/PUT routes. - Drop `MCPServerConfig.headers` from `SseServerConfig` / `StreamableHttpServerConfig` and remove the headers textarea from `ServerConfigModal`. Custom headers are now entered only in `ServerSettingsForm` and reach the wire via `settings.headers`. - Wire `ServerSettingsModal` into `App.tsx`; saves dispatch through a new `useServers.updateServerSettings` mutator. `updateServer` carries the existing settings across config-only saves so they aren't dropped. - Apply settings on the *first* outbound request after a server loads — no need to open the settings form first: - `settings.headers` → SSE / streamable-http wire headers. - `settings.metadata` → `InspectorClientOptions.defaultMetadata`, merged into every outgoing `_meta` (call-time keys win). - `settings.requestTimeout` → `InspectorClientOptions.timeout`. - `settings.connectionTimeout` → Promise.race wrapper around connect(). - `settings.oauth*` → pre-seeded OAuth client credentials. - `RemoteConnectRequest` carries `settings` so the remote backend can source transport headers from the same source as the local path. - `oauthClientSecret` is persisted in `mcp.json` alongside stdio `env` values, both protected by the file's `0o600` permission. - Update `specification/v2_servers_file.md` with the new shape and the removal of the legacy headers field. Tests: round-trip in `serverList.test.ts`; persistence on POST/PUT in `servers-route.test.ts`; first-connect header injection in `transport.test.ts`; `defaultMetadata` merging in `inspectorClient.test.ts`; `useServers.updateServerSettings` + `updateServer` settings-preservation in `useServers.test.tsx`. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(servers): rebuild InspectorClient on every connect so settings refresh The previous onToggleConnection branch reused the existing InspectorClient on a same-server reconnect, which froze settings at the moment the client was first constructed. Editing headers, metadata, timeouts, or OAuth credentials in the settings form was therefore invisible until the user either switched servers or reloaded the page — surprising right after a visible save. Always call setupClientForServer on (re)connect so the latest target.settings flow through to the transport's headers / requestInit, to defaultMetadata, and to the OAuth manager. Also adds remoteClientTransport tests that pin the wire-level contract: settings round-trip through the /api/mcp/connect body when supplied, and the field is omitted when no settings are configured. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(history): stop wiping pendingRequestEntries inside the connect event State managers (ManagedToolsState, ManagedPromptsState, ManagedResourcesState, ManagedResourceTemplatesState, ManagedRequestorTasksState) handle the client's "connect" event by calling `void this.refresh()`. The synchronous portion of refresh runs through SDK Protocol.request → transport.send → MessageTrackingTransport.send → trackRequest, all before yielding to the network — so each list request gets tracked into MessageLogState's pendingRequestEntries inside the same connect event dispatch. MessageLogState's own "connect" listener was registered after those state managers (it's the last one constructed in App.tsx) and cleared pendingRequestEntries on every connect. Listener order meant the clear ran *after* the requests were tracked but before their responses arrived, leaving the history with unmatched "response" entries marked PENDING. The clear-on-disconnect (via statusChange → disconnected) already handles session-boundary cleanup, so clear-on-connect was purely defensive and now actively harmful. Drop the connect listener; keep the disconnect path. Tests: updated the unit test to assert the new behavior (connect does not clear); added an integration regression that wires a sibling connect listener firing listTools and asserts the response folds into the request entry instead of appearing as an orphan. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * chore: drop accidentally tracked debugging screenshot pending-requests.png was a local debugging artifact attached to a conversation, not source. Remove it from the tree (the working copy is kept so the user still has it locally) and rely on the existing untracked behavior going forward. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * test(servers): add e2e check that settings.headers reach upstream MCP Stitches the full v2 pipeline together in one assertion: browser-style RemoteClientTransport → POST /api/mcp/connect (settings in body) → backend createTransportNode applies settings.headers to requestInit → SDK StreamableHTTPClientTransport → upstream MCP test server Captures the upstream HTTP request via FetchRequestLogState and asserts both `X-Tenant` and `X-Trace` are present on the outbound POST. Catches any regression that breaks the wire between the form and the upstream server in a single integration test, instead of the previous coverage which exercised each hop in isolation. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(ServerSettingsModal): invert ListToggle compact prop so icon matches state ListToggle's `compact` prop represents the list's *current* compact state: true means "currently compact" and renders the expand icon to advertise the action (open up). The settings modal was passing `compact={allExpanded}` — exactly inverted. When all sections were open it showed the expand icon (suggesting "open more") and when collapsed it showed the collapse icon, opposite of every other ListToggle caller in the app. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * style(HistoryListPanel): match ServerListScreen Export button — rename + default variant Renames "Export JSON" to "Export" and switches from the subtle ToolbarButton to the same `Button variant="default"` used by ServerListControls so the two Export affordances look identical across screens. Also moves the button to the right of the ListToggle (mirroring the order on the server list screen) and disables it when there are no entries to export. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * style(HistoryListPanel): move Clear button to top toolbar, match Export style Moves the Clear affordance out of the "History (N)" section header and up to the panel toolbar alongside Export, where it now uses the same `Button variant="default"` styling. Order in the toolbar is now: ListToggle · Clear · Export. Clear is disabled when there are no unpinned entries to remove (pinned entries are intentionally preserved). ToolbarButton helper is unused after this change and dropped. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * chore: ignore local debugging screenshots at repo root Stops `git add -A` from re-staging the `*.png` files Claude Code drops at the repo root from attached conversation screenshots. The pattern only matches the top-level (`/*.png`) so legitimate image assets in subdirectories (e.g. `clients/web/public/`) still get tracked when explicitly added. Also untracks the two screenshots that slipped through previously (initialize-request.png, pending-requests.png). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * style(LogStreamPanel): match HistoryListPanel button styling Clear / Export / Copy All on the Logging screen were rendered with the subtle ToolbarButton helper while History's equivalents use `Button variant="default"`. Switch all three to the default variant so the toolbars look identical across screens and mirror the disabling behavior — each button is now disabled when `entries.length === 0`. ToolbarButton helper is unused after this change and dropped. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(servers): address PR review — debounce save, preserve-on-omit PUT, validate shape, lift connect timeout Round-up of fixes from https://github.com/modelcontextprotocol/inspector/pull/1353#issuecomment-4526861959: 1. ServerSettingsModal saves were firing on every keystroke (one PUT + full mcp.json refresh per character). App.tsx now debounces by 300ms and flushes on modal close so a burst of edits coalesces into one PUT and the form value can't flicker back mid-typing. 2. `PUT /api/servers/:id` with an omitted `settings` field used to silently wipe the persisted settings node. The route now treats the three intents distinctly: - field omitted → preserve existing on-disk settings - explicit null → clear the settings node - settings object → validate + apply `useServers.updateServer` is simplified to drop the client-side carry-through hack, and the test that documented the broken behavior is flipped + augmented with explicit-clear and bad-shape cases. 3. `connectionTimeout` Promise.race is moved from App.tsx into `InspectorClient.connect()`, reading `this.serverSettings?.connectionTimeout`. TUI/CLI consumers now get the same behavior; on timeout the client internally disconnects so the next connect starts clean. 4. Custom Headers hint in `ServerSettingsForm` notes that `Authorization` is owned by the OAuth flow when OAuth is configured — avoids the sharp edge where the SDK's authProvider silently overwrites a user-set Authorization header. 5. New `validateSettings` on the server route rejects malformed shapes (`settings: []`, `settings.headers: "oops"`, non-numeric timeouts, etc.) with 400 instead of letting them persist and crash the UI. 7 (nit). Renamed `settings` → `savedSettings` inside `setupClientForServer` so the variable holding the persisted settings doesn't shadow the later locally-derived values. Follow-ups opened separately: #1356 — Move oauthClientSecret out of mcp.json into the OS keychain (review #6). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(servers): address PR review pass 2 — connect-timeout test, unmount cleanup, tagged intent, lenient settings read Pass-2 review at https://github.com/modelcontextprotocol/inspector/pull/1353#issuecomment-4527089705: (a) New integration test pins the new client-layer connect-timeout contract: a hanging stub transport with `connectionTimeout: 50` rejects with the descriptive error and lands the client in "error" status. Documents what TUI/CLI consumers should expect from the lifted Promise.race in InspectorClient.connect(). (b) useEffect cleanup clears `pendingSettingsTimerRef` on unmount so a route change or HMR mid-debounce can't fire one stale PUT against an unmounted component. (c) Refactor the PUT-route `settingsIntent` discriminator from `"preserve" | "clear" | { value }` to a tagged union with `kind`, and consume it with a `switch` for exhaustiveness — reads cleaner and makes adding a fourth intent painless. (d) `normalizeMcpServers` now runs `validateSettings` on the on-disk settings node and silently strips it if invalid, mirroring the existing lenient-on-read pattern from `normalizeServerType`. Closes the read/write asymmetry where a hand-edited malformed `settings` could survive on disk and get carried through preserve-on-PUT. New route test pins the strip-on-GET + no-propagate-on-preserve-PUT behavior. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(servers): pass-3 review — patch-style PUT, toast on close-flush fail Pass-3 review at https://github.com/modelcontextprotocol/inspector/pull/1353#issuecomment-4529276897. (b) `PUT /api/servers/:id` now treats both `config` and `settings` as optional patches: - field omitted → preserve the on-disk value - explicit body value → apply - settings: null → clear `updateServerSettings` stops reading `existing.config` from the in-memory `servers` snapshot, which closes the closure-capture hazard where a refresh between debounce-schedule and flush could silently revert a separate edit. The route is the single source of truth for transport config now. New tests cover settings-only PUT, rejected non-object config, and the empty-body no-op patch case. The previous "rejects a missing config" test is rewritten to document the new no-op-preserve semantics. (c) Inline comment near the PUT rebuild loop documents that writing the full mcpServers map back is what gives the route its desirable self-healing property for malformed settings on *other* entries — deliberate side-effect, not an accident of normalization. (d) `flushPendingSettings` surfaces failures via @mantine/notifications instead of swallowing them. The modal already closed by the time the flush PUT resolves, so a silent fail would leave the user thinking their last edits saved (especially painful for the OAuth client secret). The notifications wiring was already mounted in main.tsx. Pass-3 (a) — settings form being effectively uncontrolled mid-edit — is acknowledged in the reply: working as intended today; the proper fix is internal modal state when file-watching (#1345) forces the issue. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(servers): pass-4 review — strip smuggled settings out of config Pass-4 review at https://github.com/modelcontextprotocol/inspector/pull/1353#issuecomment-4529861349. Main finding: `normalizeServerType` spreads unknown keys, so a body that nested `settings` inside `config` would smuggle a settings field onto the stored entry without ever passing through `validateSettings`. Strip the settings key off the incoming config in `buildStoredEntry` so `validateSettings` remains the single write path for the settings node — the "one source of truth" invariant pass 2 set up holds again. Two new tests pin the strip on both routes: - POST with `config.settings = { bogus }` lands an entry with no settings node on disk. - PUT with `config.settings = { bogus }` + `settings: null` clears the real settings *and* doesn't re-attach the bogus payload via the spread. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(servers): pass-5 review — absorb late connect rejection, tighten validateSettings, log smuggle warning Pass-5 review at https://github.com/modelcontextprotocol/inspector/pull/1353#issuecomment-4529954490. Main finding: when the connect-time `Promise.race` lost to the timeout, the original `this.client.connect(this.transport)` promise was left pending. After `disconnect()` tore the transport down, real transports (SSE / streamable-http) would reject that promise later — with no handler attached, producing a Node `unhandledRejection` (and a vitest suite warning). Attach a no-op `.catch` right after creating the connect promise so Node sees the late rejection as handled. Nit (a): `validateSettings` was casting the raw object through to `InspectorServerSettings`, which let unknown keys ride along onto disk. Replaced the cast with an explicit pick-and-build over the known fields. Unknown stowaways now silently drop on the way through. New route test pins the behavior. Nit (b): `buildStoredEntry` now logs a `warn` via the route's pino logger when an incoming `config` carries a `settings` key. The strip is still defensive (it isn't a documented contract) but the warning lets a misconfigured client find out about the wasted field instead of debugging a silent drop. Nit (c) — the defensive `?? {}` fallback in `buildStoredEntry` — acknowledged in the review reply; left in place as belt-and-braces. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(servers): pass-6 polish — scope absorber, thread id, drop empty OAuth, guard `in` Pass-6 review at https://github.com/modelcontextprotocol/inspector/pull/1353#issuecomment-4529979135. Reviewer signed off as ready; landing the four marginal cleanups since they're all small. (a) `connectPromise.catch(() => {})` moves inside the `if (connectTimeoutMs > 0)` branch. The await-only path propagates rejection cleanly through the outer try/catch — the absorber is only needed when the race leaves the original promise unattended. Code now reads as "absorber is tied to the race," which is the real invariant. (b) Thread `id` (POST id, PUT newId) through `buildStoredEntry` and include it in the smuggle warning's pino bindings. Two clients smuggling simultaneously are distinguishable in the log instead of indistinguishable. (c) `validateSettings` coerces empty-string OAuth fields to absent. The form emits `""` when the user clears an input, and an empty `oauthClientId` on disk would later be misread as "OAuth configured." New route test pins it. (d) `buildStoredEntry` adds a null/object guard before `"settings" in configObj`. The route layer pre-checks this on POST and PUT, but the helper's `unknown` parameter is a weak contract; the guard makes it safe to call from anywhere. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(spec): reflect settings persistence + patch-PUT contract in v2_servers_file The doc was authored when `mcp.json` only stored the server list. Bring it forward to match what shipped in #1352 / PR #1353: - Summary calls out that the file also stores per-server settings and links into the Per-server settings section. - On-disk format note explains that `settings` is an optional Inspector- specific extension that other MCP tools ignore. - New top paragraph on the Per-server settings section explains the basic-config / settings UI split rationale (simpler form for the common case, advanced settings tucked behind a second dialog). - UI bullet is rewritten to match the patch-style PUT semantics that emerged across the review cycle: settings-only and config-only PUTs preserve their counterpart inside the write lock, so neither modal can silently clobber the other half. The earlier "must re-send settings on PUT" hack is gone. - New bullets capture the save cadence (300ms debounce + flush-on-close + failure toast), the patch-PUT contract (omit-preserve / null-clear / rename), and the write-path / read-path validator gates. - Secret storage bullet links to #1356 (OS-keychain follow-up). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(spec): show a third entry with the optional settings node The On-disk format example demonstrated only basic stdio entries. Add a streamable-http entry with a populated `settings` node so readers can see the shape inline rather than having to scroll to the Per-server settings section. Inline comment in the block cross-links there for the full contract and notes that other MCP tools ignore the extension key. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>	2026-05-24 19:23:26 -04:00
Cliff Hall	8bb389c8e3	test: switch integration project to folder-based manifest (#1314) (#1318) ... * test: switch integration project to folder-based manifest (#1314) Replace the `integrationTests` enumeration in vite.config.ts with a folder glob (`src/test/integration/**/*.test.ts`). Files under that folder run in the integration vitest project (node env, 30s timeouts); everything else runs in the unit project (happy-dom). This removes the silent misclassification trap from #1307: a new integration-style test no longer falls into the unit project (and happy-dom) just because someone forgot to add it to the array. The 13 ported v1.5 tests + their helpers move under src/test/integration/ in a strict mirror of core/ source layout (mcp/, auth/, auth/node/, storage/). Renames use git mv so history is preserved. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * test: address PR review on #1318 (glob extensions + AGENTS doc) - vite.config.ts: widen integrationGlob from `*.test.ts` to `*.test.{ts,tsx}` so the integration folder fully owns its contents. Without this, a `.test.tsx` placed under src/test/integration/ would match the unit project's `{ts,tsx}` include, fail the unit `.ts`-only exclude, and silently land under happy-dom — the exact misclassification trap this PR set out to close. - AGENTS.md: update the line describing where integration tests live; it still pointed at the old `src/test/core/` location and the removed `integrationTests` enumeration. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>	2026-05-18 11:54:49 -04:00

Renamed from clients/web/src/test/core/inspectorClient.test.ts (Previous history)