Four findings, all real.
- No catalog-level bound. The per-skill caps bound what ONE entry can
cost; nothing bounded how many entries there are, and SEP-2640 puts no
ceiling on a catalog — every entry costs at least one resources/read,
so a large listing made `--verify` run indefinitely and transfer
unboundedly. Adds SKILL_MAX_CATALOG_SKILLS / SKILL_MAX_CATALOG_BYTES.
Entries past the budget are still reported, with their static findings
and an `incomplete` reason, rather than dropped or failed: they were
not checked, which is neither a pass nor a verdict against the server.
- `responseBytes` charged `text.length`, which counts UTF-16 code units.
That undercharges non-ASCII by up to 3x, so a decoy block of emoji kept
the counter under 16 MiB while the wire carried far more. Adds
`utf8Length`, an allocation-free UTF-8 byte count — deliberately not
TextEncoder, which would copy a payload the server chose the size of.
- `checkSkillNameCollisions` was O(N^2) in both work and output for a
group of N: every entry filtered all N URIs and embedded the other
N-1. Duplicate names are legal and a server controls N. Now names a
bounded sample and counts the rest.
- `--verify` help documented exit 7 but not exit 8.
The three-tier bounding (per skill, per skill on the wire, per run) is
now a table in clients/cli/README.md, since the run bound is this tool's
limit rather than the spec's and should not read as a conformance rule.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ND46Ci64DC3tzBy9aG81Nc
Signed-off-by: cliffhall <cliff@futurescale.com>
Four of the five findings share one root cause: the tri-state `outcome`
added in round 15 was not propagated to consumers, which kept branching
on `ok` — true for an `incomplete` report, since nothing that was checked
was wrong.
- CLI `summarizeSkillVerification` counts off `outcome`, not `ok`, so a
truncated walk no longer prints "no conformance errors" one line before
exiting SKILL_INCOMPLETE. A mixed failed/incomplete catalog reports
both counts rather than letting the louder verdict hide the quieter.
- `verifySkills` sets the truncation reason only when entries were
actually left unread. A file crossing the byte budget as the FINAL
entry stopped nothing, and "Stopped after 3 of 3" both read as a
contradiction and demoted a fully-read skill out of `verified`.
- The TUI status line switches on `outcome` through a `Record` over the
union, so the INCOMPLETE arm is reachable and a fourth outcome would
be a type error rather than a silently missing label.
- The `incomplete` doc paragraph said it makes `ok` false. It does not,
deliberately — corrected, and it now names `outcome` as the thing a
consumer must branch on.
- The web sidebar composes in the per-URI `duplicate-name` warning, so
two colliding skills are badged in the catalog instead of looking
clean until one is selected.
The suppressed `selfAttempted` finding is stale — fixed in round 16.
The existing TUI INCOMPLETE test was passing for the wrong reason: its
fixture understated every size, which is itself a mismatch, so the report
was `failed` and only the old `ok`-first branch printed INCOMPLETE. It is
rebuilt with honest digests and sizes, so truncation comes from the
declared-size prefilter and the report is genuinely incomplete.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ND46Ci64DC3tzBy9aG81Nc
Signed-off-by: cliffhall <cliff@futurescale.com>
**Verification is a tri-state now, because two states could not be right.**
Round 12 was correct that `ok: true` for a manifest whose unread 513th file
may be tampered with is a false pass. Round 15 is correct that forcing `ok:
false` calls a server nonconformant for exceeding limits SEP-2640 states as
SHOULD NOT, with hosts free to support more — contradicting this module's own
rule that a warning never fails a report. Both hold, so:
verified -> exit 0 everything checked, everything passed
failed -> exit 7 a MUST was broken
incomplete -> exit 8 nothing checked was wrong; the walk was cut short
`ok` keeps its narrow meaning, `allSkillsVerified` is stricter than
`every(r => r.ok)` since it selects the exit code, and `anySkillFailed`
separates 7 from 8. A job that tolerates oversized catalogs can allow 8 and
still fail on 7.
**`manifestKey` still used `JSON.stringify` on the entry** — the same crash I
fixed in the TUI in round 12, one file over.
**Directory children were navigable outside the skill root.** A server could
return `skill://other-skill/...` and clicking it left the tree, with "Up"
only comparing equality against the root. Containment is checked on the
normalized URI now, and an outside child renders as text rather than a link.
Worth recording from the tests: `..` cannot escape the authority, so
`skill://a/nested/../x.md` resolves back inside and must stay navigable.
**The TUI matched report rows by raw URI**, which missed a row recorded under
an equivalent spelling while `extraReportFiles` suppressed it as covered.
Normalized, like the membership test beside it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ND46Ci64DC3tzBy9aG81Nc
Signed-off-by: cliffhall <cliff@futurescale.com>
Round 10 raised no new inline comments; three suppressed findings, two fixed
and one declined with a reason.
**`Object.is` held `0` and `-0` distinct**, so a listing carrying JSON `0`
against a served YAML `-0` produced a mismatch — reported as "the listing
says 0 but the served SKILL.md says 0", a false finding with an
unintelligible explanation. JSON does not distinguish them, so neither may
this: `===` for finite numbers, `Object.is` kept only for the non-finite
cases it was introduced for.
**Browsing a supporting file erased an observed failure.** The frontmatter
check ran off whatever the viewer was showing, so opening a second file made
`showingSkillMd` false and dropped the finding and its error count, with the
skill unchanged. The entry's text is skill-scoped now, written by whichever
read produces it first — the on-selection preview or a verification — and
invalidated only with `manifestKey`. That also subsumes the earlier
same-fetch requirement rather than sitting beside it.
**Declined: the second stderr line on exit 7.** `--verify` writes its summary
and then the ordinary `ErrorEnvelope` that EVERY non-zero exit writes, which
is the documented contract and exactly what `--strict` does. Suppressing it
would make this one command's failure output unparseable by a caller
branching on `.code`. What was wrong is the README, which implied stderr
carries only the summary; it now states both lines and why.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ND46Ci64DC3tzBy9aG81Nc
Signed-off-by: cliffhall <cliff@futurescale.com>
Four findings.
**The CLI dropped what it promised to print.** `getSkill` transforms the
`skills/get` envelope to a `SkillEntry`, so every other member the loose
schema accepted was gone before the CLI saw it — including the `ttlMs` /
`cacheScope` SEP-2640 explicitly leaves open. The code contradicted a comment
two lines above it saying a CLI whose contract is "print the result" must not
reshape one. `getSkillResult` returns the envelope whole; `getSkill` is a
one-line unwrap over it, which matches the callers: the UIs want the entry,
the CLI wants the result.
**The frontmatter verdict and the digest came from different reads.** The
preview fetch and the Verify fetch are separate `resources/read` calls, so a
resource changing between them could pair a verified digest with a
frontmatter verdict for other bytes. `VerificationState.entryText` now holds
the SKILL.md as the verification read it, written in the SAME state update as
the verdict — a separate write was clobbered, because `write` returns a fresh
object and drops anything not named in it. The preview stays the fallback so
a reader who has not clicked Verify still gets the check.
**The Conformance badge excluded the frontmatter findings** it renders, so it
said `0 error(s)` above a red mismatch. Digest and size mismatches stay in
their own badge: "the listing is wrong" and "the bytes are wrong" are
different answers.
**A repeated skill URI collided two React keys** in the TUI list, in a pane
whose job is to show both entries. Index-keyed like the manifest rows.
The web fixture also gains its collision pair in the served-file lookup —
without it they were handed another skill's SKILL.md, which the badge fix
correctly began reporting as a frontmatter error.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ND46Ci64DC3tzBy9aG81Nc
Signed-off-by: cliffhall <cliff@futurescale.com>
Five findings, all real.
**Critical — `verifySkills` selected `contents[0]`.** These bytes are hashed
against the file's advertised digest, so accepting a block the server
labelled something else verifies one file's content against another file's
digest, and can report that as `verified`. A false pass is worse than a
missing check. `contentsFor(result, uri)` now selects by NORMALIZED identity,
which keeps the canonicalized-echo case that motivated the original code
while rejecting an unrelated block, and reports a read error when nothing
answers for the URI. This is what `onReadSkillFile` already did; the two
paths hash bytes against digests and must not disagree about which bytes.
**`skills/get` was dispatched without the extension gate** that `skills/list`
had. Hoisted into `assertSkillsSupported` so the two cannot drift — declaring
the extension commits a server to both. Without it, an undeclared server's
-32601 is indistinguishable to a script from the -32602 a declared server
returns for a URI it does not serve.
**The TUI could strand a user on a hidden Skills tab.** The tab left the bar
when the gate went false but `activeTab` did not, so the pane kept rendering
for a server that never declared the extension. Reset to `info`, following
the Auth precedent — additionally gated on `connected`, because this gate
reads a server declaration and would otherwise fire during a reconnect.
**The TUI keyed a verdict by URI**, so a refresh replacing the manifest under
the same URI left hashes computed for the previous snapshot describing the
new one. Keyed on the entry now, as the web screen already was.
**A failed "Load more" discarded the pages on screen and the retry cursor.**
Both are preserved now, and the success and failure paths share one
staleness-guarded `commit` helper so they cannot drift on which results they
may write — which is how they came to disagree.
Two nits: the six-skill counts. The docs sentence was more wrong than
reported (five awkward skills of six, three outright violations) and is
rewritten rather than renumbered.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ND46Ci64DC3tzBy9aG81Nc
Signed-off-by: cliffhall <cliff@futurescale.com>
Closes#2248.
Completes SEP-2640 support across all three clients, and closes the one
obligation #2234 deliberately left open.
- `checkSkillFrontmatterMatch` compares a served SKILL.md's own YAML
frontmatter against the entry the listing advertised, field by field. No
digest can cover this: a digest is taken over the bytes the server served,
so it proves the file was not altered in transit and says nothing about
whether the listing described it honestly. Needs a real YAML parser; `yaml`
was already a root dependency, so this adds no package — but it does newly
put it on core/'s import graph, hence the three bundler `external` lists.
- `resources/directory/read`: result schemas defined against the normative
text, `InspectorClient.readResourceDirectory`, a Directory section on the
Skills screen, and `--method resources/directory/read` in the CLI. The
client refuses the call locally when the server did not declare
`directoryRead`, which is the SEP's MUST NOT. A child the directory lists
but the entry does not is marked `not listed` rather than merged into the
manifest — the SEP calls a directory result a live observation and forbids
treating it as extending the manifest.
- CLI: `skills/list`, `skills/get` and `resources/directory/read`, plus
`--verify` — one NDJSON report per skill, a summary on stderr, exit 7 on a
violation. Reads are sequential; a conforming manifest may declare 512
entries.
- TUI: a Skills pane, shown only when the server declares the extension. Each
row carries its conformance verdict as a glyph as well as a colour, since
the pane is read over ssh and through script(1). Enter verifies.
- `verifySkills` re-throws `AuthRecoveryRequiredError` rather than recording
it per file: it says the session's authorization expired, so absorbing it
would report N identical read failures and swallow the error the TUI and
the web commands key off to reauthorize.
Two open questions settled, both in code comments where they will be found:
`skills/get` carries no caching attributes because SEP-2640 leaves the
question open in as many words, so requiring them would fail a conforming
server; and there is no `PagedSkillsState` because every consumer of this
list is a whole-catalog verdict — computed over page one of three, "this
server's skills conform" is not merely partial but wrong.
The fixture grows two skills, both for checks that were otherwise
undemonstrable: `lying-listing` (listing and file disagree, digest still
verifies) and `stale-manifest` (serves a file its manifest does not declare).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ND46Ci64DC3tzBy9aG81Nc
Signed-off-by: cliffhall <cliff@futurescale.com>
Adds `docs/cli-smoke-testing.md` — a copyable connect → list → call →
assert workflow for testing an arbitrary MCP server from a shell or a CI
job, plus index rows in the root README and a pointer from the CLI
README.
The gap this fills is a user-facing one. `clients/cli/README.md` is a
flag reference and `scripts/smoke-cli.mjs` is an internal end-to-end
check of the launcher → CLI path; neither tells someone how to compose
those flags into a check they run on every commit against their own
server.
Covers, in order: connecting over stdio/HTTP/SSE (including the `--`
separator, which splits the opposite way under `--cli`); `--format json`
and its `{result[, appInfo]}` envelope; asserting a tool exists with
`jq -e`; one representative `tools/call` and the `--tool-arg` vs
`--tool-args-json` coercion difference; the exit-code map and how to
branch on it; keeping OAuth non-interactive with `--stored-auth-only`
and an isolated `MCP_STORAGE_DIR`; and negative assertions — refusals,
a secret-shape scan of captured output, and `--strict` schema
portability. Ends with a complete script and a GitHub Actions job.
Every command and output shape in the guide was verified against the
bundled composable test servers over both stdio and HTTP rather than
transcribed from the reference, which corrected three claims that would
otherwise have shipped wrong:
- `--tool-arg zip=012` does not send `12`. `012` is not valid JSON, so
it is sent as the literal string. The real trap is `zip=10001`, which
does coerce to a number and is then rejected by a string schema.
- `if ! cmd; then status=$?` yields `0`, because `!` inverts the
pipeline's status — every failure class would look identical. The
guide uses `|| status=$?` and says why.
- The sample script does not report a failure class on its own; that is
the opt-in `case` block from §5.
The §8 script was run end to end against a live HTTP test server, and
exits 0/3/4/5/6 were each reproduced against a real server.
Closes#1886
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KsKC3HniVF9dCZtDVbYp5B
Signed-off-by: cliffhall <cliff@futurescale.com>
`clients/cli` resolved `@types/node` 24.13.1 against the root's 24.13.3.
Neither client declares it — #2196 consolidated it to the root — so the cli
copy arrived transitively via `@types/express` and was constrained by no
range of ours. `verify:dep-lockstep` reported OK and was right to: it
compares only packages that reach one `tsc` **program** from two installs,
and no one program sees both copies.
Fixes the instance and the class, which the issue notes are not exclusive.
**The class.** The guard grows a second tier alongside the program-derived
one, asking the weaker but broader question `AGENTS.md`'s "one version per
install-crossing dependency" rule actually states: does a package this repo
declares anywhere resolve to two versions across our installs at all? The
candidate set is every name in any install's `dependencies`/`devDependencies`
that more than one install holds a **top-level** copy of — 17 packages today.
Nested copies are excluded: one exists because some dependency asked for a
different version, so it is that dependency's range, not ours.
Neither tier subsumes the other. The program tier sees a copy no manifest
names (the `@modelcontextprotocol/sdk` case, arriving through another
package's `.d.ts`). The declared tier sees a transitive copy no program
loads (this issue), two *clients* disagreeing with no root copy involved
(`@types/react`, web against tui), and the peer shadows `AGENTS.md` calls out
as ungated — `eslint`, `typescript`, `vitest`.
Deny-by-default, with its own empty `TOLERATED_DECLARED_SKEW` and the same
within-a-major rule, now stated once in a shared `toleratesSkew` so the two
tiers cannot drift.
**The instance, and three others the new tier surfaced.** All four were
patch-level:
- `@types/node` — cli 24.13.1 vs root 24.13.3. Transitive, so it takes an
`overrides` entry in `clients/cli` (the repo's sanctioned lever for a
transitive pin), at the root's own `^24.12.4` rather than a frozen number.
- `@types/react` — web 19.2.18 vs tui 19.2.17.
- `react` — web 19.2.8 vs root/tui 19.2.7.
- `@vitejs/plugin-react` — web 6.0.5 vs root 6.0.2.
The last three are `npm update` in the stale install: it moves the lockfile
within the declared range without widening a range, which matters for the
root `react ^19.0.0` that `ink`'s externalized-bundle exemption depends on.
No declaration or range changed.
Tests: 4 end-to-end cases (the fixture gains manifest deps and a client-side
`solo` install, so the skew is real and no program holds both copies) and 13
helper cases.
Closes#2226
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01J86a6vMcA9csPtdsg7ZcCZ
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot round 3 on #2256: both detectors were one-sided. Each measured
the sub-floor remainder against the real clock, so a preempted worker
overshoots the deadline, the remainder goes negative, and the unfixed
`remainingMs <= 0` bound takes the same branch and prints the same
message — the test passes on an implementation with no floor. It could
never fail wrongly, but it could silently stop testing anything, which on
a flake fix is the wrong half of that guarantee to keep.
Both now stub `performance.now()`: the web case advances it inside the
fetch so the second grant's remainder is exactly
MIN_REVOCATION_REQUEST_BUDGET_MS - 1, and the CLI case freezes it so the
remainder at the check is the budget itself. Positive and under the floor
on every machine, which is the only state that separates the two bounds.
Verified as detectors with the clock stubbed: reverting the core floor
gives 1 failed / 57 passed, reverting the CLI floor 1 failed / 14 passed.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017wEXtbs8UEHUMxDEAxbs99
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot round 1 on #2256: the `budgetMs: 0` case passes under the old
`remainingMs <= 0` bound too, so nothing in the CLI suite detected the
floor added to `sendPlans`.
The new case gives it a budget that is positive and below the floor, and
asserts on *this loop's* exhaustion message — the one naming the server
URL. That distinction is the test: with the floor removed the plan is
handed a 4ms budget and core's identical floor declines it one level
down, so a bare "budget was exhausted" match passes either way. Verified
as a detector — 1 failed / 14 passed with the CLI floor reverted, 15
passed with it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017wEXtbs8UEHUMxDEAxbs99
Signed-off-by: cliffhall <cliff@futurescale.com>
The shared teardown deadline was decided on `remainingMs > 0` against
`Date.now()`. Both halves of that are noise-sensitive: the deadline is
enforced by a `setTimeout`, and a wall clock at millisecond resolution
can still read a hair short of it when the loop comes round, so the next
grant inherits a fractional budget and spends it on a request that cannot
possibly complete. The same run then issues one request or two depending
on scheduling — which is what makes the "shares one deadline across
grants" test intermittent on CI (#2252).
Measure the deadline with `performance.now()` (monotonic, so an NTP step
or a suspend cannot expire or extend the budget, and sub-millisecond, so
none is lost to rounding), and skip any grant reaching the loop with less
than MIN_REVOCATION_REQUEST_BUDGET_MS left. The skipped grant is already
reported as `failed`, so nothing is silently dropped — the needless call
to the authorization server is.
The CLI's outer per-plan budget in `sendPlans` shares both the shape and
the defect, so it gets the same treatment.
One consequence worth naming: a monotonic clock hands `revokeToken` a
fractional budget, and Node's `AbortSignal.timeout` throws
`ERR_OUT_OF_RANGE` on a non-integer delay — before the fetch, so the
request would never be sent and the revocation would report that as its
failure. `revokeToken` now rounds its own budget to whole milliseconds,
which also covers any other caller passing a fractional one.
Tests: the epsilon skip is pinned by a grant that lands inside the floor
but above zero (removing the floor fails it, verified); the fractional
budget by a `revokeToken` call at 19.996ms.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017wEXtbs8UEHUMxDEAxbs99
Signed-off-by: cliffhall <cliff@futurescale.com>
- Complete the bundler external lists against the AGENTS.md invariant rather
than against today's reachability: chokidar, hono and react to clients/cli,
chokidar and hono to clients/tui, react to the web runner. Every root-declared
package core/ imports at runtime is now named in all three, so a new import in
core/ cannot silently inline one.
- AGENTS.md: split the single over-broad bullet in two. commander, open,
@hono/node-server, vite and @vitejs/plugin-react are root dependencies reached
only from client code, so they are not part of the core/ set and need naming
only in the external list of the client that imports them. Give the
vitest.shared.mts pins their own rule, stating the react/react-dom exception
instead of implying every root-owned package is aliased to the root.
- vitest.shared.mts: the umbrella comment now says root-OWNED rather than root
runtime dependency — express is test-only and lives in the root
devDependencies — and notes that the list is deliberately not the complete
root set, since a package that resolves unambiguously needs no pin.
- local-dev skill: document why react/react-dom stay client-pinned, which the
previous text contradicted.
npm run local:gate and npm run verify:bundle-externals pass.
Signed-off-by: cliffhall <cliff@futurescale.com>
Every runtime dependency core/ imports is now declared once, in the repo-root
package.json, and nowhere else. Deleting the client-side copies does not merely
align them: a package installs only into an install root that declares it, so
with no second declaration a second copy cannot be created at all — the failure
class #1896 was opened for and #1965/#2029 built verify:dep-lockstep to detect.
- clients/{web,cli,tui,launcher}: dropped ajv, atomically, chokidar, commander,
open, pino, zod, @napi-rs/keyring, hono, @hono/node-server and react, as each
applied. clients/cli and clients/launcher now declare no runtime dependencies.
- Root hono floor raised ^4.12.18 -> ^4.13.1, keeping the higher of the two
ranges that were in the tree rather than silently lowering web's.
- tsup external lists: added ajv and zod to all three bundles, plus atomically
and open (cli) and atomically (tui). tsup externalizes only what the nearest
manifest declares, so a package that moves to the root is inlined unless it is
named explicitly.
- vitest.shared.mts: repointed the pino, hono, @hono/node-server, atomically,
chokidar and @napi-rs/keyring pins from <client>/node_modules to the repo
root, joining express, yaml and proper-lockfile. Left behind they would
resolve to a directory that no longer exists, or to a transitive copy an
unrelated dependency dragged in. react and react-dom stay client-pinned so
the pair stays version-matched within an install.
- Docs: AGENTS.md dependency placement, the root README and the local-dev skill
describe the consolidated layout.
verify:dep-lockstep candidate set: 4 -> 2 (hono and pino fell out; only
@standard-schema/spec and zod still reach one tsc program from two installs).
npm run local:gate and npm run pack:verify both pass.
Signed-off-by: cliffhall <cliff@futurescale.com>
Declare the toolchain every client uses once, at the repo root, and delete the
per-client copies: eslint, @eslint/js, typescript-eslint, globals, prettier and
typescript (all five manifests), plus vitest, @vitest/coverage-v8 and
@types/node (four). Adds the latter three to the root manifest; the first six
were already there.
Every client reaches the root copy by walk-up — npm run puts each ancestor
node_modules/.bin on PATH, and Node and TypeScript resolve modules and @types
the same way — so nothing about the client scripts changes. clients/launcher now
declares no devDependencies at all and its validate is unchanged.
The copies had already drifted: globals sat at ^17.7.0 at the root against
^17.4.0 in all four clients, and typescript-eslint at ^8.65.0 against ^8.56.1.
Consolidating takes the root's version in both cases, and makes the exact
prettier pin (#1790) a single pin rather than five that can diverge.
Tools only one client runs stay where they are — tsup, tsx, vite, vite-node,
playwright, storybook, happy-dom, ink-testing-library, and each client's own
@types/*.
Documents the rule in AGENTS.md's Dependency placement section and the reasoning
in the local-dev skill.
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot reported "no new comments" again, with four real findings in its
suppressed block.
- The web in-flight lock was global, so a pending clear for server A silently
swallowed a click on server B's — even though the callback explicitly
supports clearing a non-active server. Keyed by server id now.
- `else if (!isActive)` skipped the resume-snapshot cleanup when the target IS
active but has no client yet, which is reachable while a session is being
built or torn down. The clear then emptied storage and left stale OAuth
recovery state pointing at credentials that no longer existed.
- `revocationAuthMethods` returned `[]` for an omitted
`revocation_endpoint_auth_methods_supported`, which does not enforce RFC
8414's default: an empty list means "the metadata said nothing" to
`selectClientAuthMethod`, which then honors the registration's own
`token_endpoint_auth_method`. The default is named literally now.
- The `v8 ignore` on the CLI budget branch was wrong, and the challenge to it
was right: an injected `budgetMs` of zero reaches it directly, through a seam
I had added myself the round before. Removed, and covered by a test that
asserts no request is made and the unattempted key is reported.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 19.
- Parsing the preconfigured client registration sat outside the per-slot
salvage, so one malformed fallback aborted the whole plan and grants carrying
their own valid credentials were never revoked. It is a *fallback*; a bad one
is now recorded and skipped, matching every other slot.
- `sendPlans` preserved an earlier success when a later key was never
attempted, contradicting the failure-first rule two lines below — the
unattempted key's grant may still be live. That branch turns out to be a
bound rather than a path (`withDeadline` fails any plan that reaches the
deadline, so a preceding plan cannot both succeed and leave zero budget), so
it carries a justified `v8 ignore` saying exactly that.
- `clearStoredAuthForRelogin` takes an optional `budgetMs`, which is what lets
the shared-budget behavior be tested without a five-second test.
- Three stale references to `planOAuthRevocation`, which no longer exists —
two JSDoc links and the `AGENTS.md` entry, which additionally still told
callers to make a separate `storage.clear()` call. Following that text would
have reintroduced the second clear the atomic step exists to remove.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 18. Three of its findings were the same check-then-act:
`planOAuthRevocation` is asynchronous, so the gap between its reads and the
following `storage.clear()` is still a window — shorter than the network one
closed in round 13, but the same shape. An OAuth completion landing at any of
those awaits saved a grant the clear then destroyed.
`OAuthStorage.takeRevocationSnapshot(serverUrl)` reads what revocation needs
and clears the server in ONE synchronous pass over the in-memory state, with
the persist after the mutation. `clearAndPlanRevocation` replaces
`planOAuthRevocation`, and there is no separate `clear` call left in any
caller — the clear cannot be forgotten or reordered.
The snapshot returns values unparsed, deliberately: schema validation is pure
and belongs after the mutation, since running it inside would reintroduce the
`await` this exists to remove.
Two consequences worth naming. A failure of the take-and-clear now REJECTS
rather than returning an outcome, because the state was not cleared and the
caller's contract is that it was; both clients already have a rejection path.
And cross-process atomicity is explicitly not claimed — nothing in this store
takes a lock, every mutation is a read-modify-write over a loaded snapshot, and
`clear` alone always had that property.
Also from round 18:
- The CLI's budget check ran before considering `plan.outcome`, so a key that
needed no network was reported as budget-exhausted — warning that a grant may
be live when that key held none.
- A disconnect failure after a successful clear was reported as "could not
clear OAuth state" in both clients. It now goes to the disconnect channel,
and the success is still reported.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 17.
`resolveClientInformation` mirrored `BaseOAuthClientProvider`, preferring the
preregistered entry. That order answers "who should I authenticate as now";
revocation asks "who minted this token", and the two diverge — the store lets a
preregistered client and an issuer-bound dynamic registration coexist, so after
a server is switched from DCR to a configured `oauth.clientId` an older DCR
grant was revoked with the configured client's credentials. RFC 7009 §2.2
answers 200 for a token the server does not recognise as the caller's, so that
reported `revoked` while the grant stayed live and the local record was already
gone — the worst available outcome.
The registration bound to the grant's issuer now wins, with the preregistered
entry as the fallback, which covers both directions: a token minted with the
configured client leaves the issuer slot empty, since the SDK writes it only
after DCR.
Stated in the comment rather than implied: this is best available evidence, not
proof. The store records client information per issuer, not per token, so
binding identity at save time remains the real fix and is a storage-shape
change beyond this PR.
Also fixed a CLI test comment still describing the pre-round-13 ordering.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 15.
The issuer check still failed open for the legacy unkeyed grant. "Unkeyed"
means its authorization server is UNKNOWN, not that the cached endpoint is
proven to own it: server metadata is a single slot a later discovery
overwrites, while a pre-SEP-2352 token survives until the first issuer-stamped
save. So a clear could post an old bearer credential to an authorization server
that never minted it.
Such a grant is now cleared without being revoked, and reported as such with a
message saying re-authorizing makes future clears revocable. The cost is that
a pre-issuer-binding entry gets no revocation — which is exactly the behavior
it had before this feature existed, so nothing regresses — and it earns
revocation back on the next authorization. Not sending is the only answer that
cannot be wrong.
Every seed in the suite is now issuer-bound, which is what a real flow writes;
the two tests that were about the legacy grant now assert the refusal, since
that is the behavior worth pinning.
Also renamed the CLI test that still described the pre-round-13 ordering.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 14. Two code fixes and a documentation sweep.
- `buildRevocationRequest` ran outside `revokeToken`'s best-effort try.
`encodeURIComponent` throws on a lone UTF-16 surrogate — valid JSON, so it
can reach there from a persisted client id or secret — and every caller has
already cleared its local state by then, so a rejection instead of a `failed`
outcome would break the documented guarantee. Construction is inside the try
now, and the catch reports `params.endpoint` rather than a `url` that may not
exist on that path.
- The fixture's Basic decode turned a malformed percent escape into an Express
500, reporting a bad credential as a server fault rather than as the
`invalid_client` 401 the endpoint means. Decoding failure is failed
authentication.
- Ten places still said revocation happens *before* the clear, which the
reorder in the previous commit made false: the settings-form description, the
`--no-revoke` help text, both READMEs, `docs/mcp-server-configuration.md`,
three JSDoc blocks, and the PR body. Swept, and the ordering is stated where
it is load-bearing rather than asserted in passing. Also merged the two
adjacent JSDoc blocks on `InspectorClient.clearOAuthTokens`.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 13. Three of its four findings were one root cause, and
it is a design flaw in the premise this PR started from.
Revoking *before* the clear meant the clear waited up to five seconds on the
network — and anything that wrote OAuth state in that window was then deleted
by a clear reasoning about the grant it replaced. A same-server reauthorization
in the TUI (whose Disconnect/Connect accelerators stay live), a stored-only web
clear for a server that becomes active mid-flight, and another CLI/TUI process
writing to the shared file-backed store all hit it.
`revokeStoredOAuthTokens` is therefore split into `planOAuthRevocation` and
`executeOAuthRevocation`, and every caller runs plan → `storage.clear()` →
execute. The snapshot still precedes the clear, because the store is where the
token, the credentials and the endpoint live; but the clear no longer waits on
anything. There is deliberately no single-call wrapper left, so no caller can
reintroduce the racy order by accident.
Fourth finding, same file: `handleConnect` now retires `clearOAuthAttemptRef`
alongside `disconnectAttemptRef`. The server-name check does not cover
disconnect/reconnect to the SAME server — the name still matches on the other
side of it — so a clear still in flight could tear down the session the connect
had just established.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 7.
- The five-second timeout was per request while grants are revoked
sequentially, so a server with N issuer slots could block the disconnect for
N x 5s — the "short timeout" bounded a request rather than anything the user
feels. It is now one budget for the whole call: each grant gets the remaining
time, and a grant reached after it is exhausted is reported as never
attempted rather than starting a fresh five seconds. `--relogin` shares one
budget across its two key spellings for the same reason.
- `AuthTab` called the same settle handler on rejection, so a failed local
clear or disconnect announced "OAuth state cleared". A rejection is not a
revocation failure — those come back as outcomes — so it now has its own
path and reports the real error. The completion is also scoped: a clear
carries the server it started on and an attempt number, and a switch retires
it, so server A's confirmation cannot appear under B.
- `clearOAuthAttemptRef` is now retired by the server-switch effect alongside
`disconnectAttemptRef`; without that an A -> B -> A round trip left both the
token and the captured name matching again. `handleClearOAuth` also
revalidates after `await disconnectInspector()`, since a switch during that
second await would have put the revision bump on the new selection.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 5.
- Deduplicating grants by token string alone could conflate two grants: a token
is only meaningful to the authorization server that minted it, so two issuers
minting the same opaque value are two grants. The second was dropped before
the issuer-mismatch check could report it, and `clear` then deleted it. The
key is now issuer+token; the ctx-less read stays suppressed only when its
token already came from a slot, which is exactly the active-issuer duplicate
and never the legacy one.
- A read failure in one issuer slot aborted `collectGrants`, so one corrupt
slot left every other revocable grant untouched while the clear deleted them
all. Each slot is now read independently and its failure carried as an
outcome alongside the grants that could still be revoked; the no-metadata and
no-endpoint returns carry them too, so a failure is never swallowed by a skip.
- The CLI no longer deduplicates its two key spellings. That could only be done
by pre-reading a single token, and `revokeStoredOAuthTokens` now enumerates
every issuer slot under a key — so a shared active token skipped the second
key entirely, taking any other issuer-bound grant under it with the local
delete. A duplicate RFC 7009 request is harmless (§2.2 makes an unknown token
a success); a missed one is the leak.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
Three review findings.
Probe responses the loop skips — and the original failed response the
substitution replaces — were dropped without releasing their bodies. On
Node/undici that holds the connection, and this loop can run on every OAuth
attempt, so repeated discovery against a 4xx candidate could exhaust the
origin's pool. Both paths now cancel, the same discipline
`core/mcp/node/authChallengeFetch.ts` uses for a discarded 401.
`fetch` follows redirects, so the candidate URL is where the probe was aimed,
not necessarily where the document came from. The source header and the console
warning now report `probe.url`, falling back to the candidate for a synthesized
response that carries none.
The CLI's stored-token refresh built its fetch from the global rather than from
`createProxyFetch()`, so a server reachable only through `HTTPS_PROXY` was
probed directly. It now builds one proxy-aware fetch and hands it to both the
discovery probe and `refreshAuthorization`, so neither leg bypasses the proxy —
this function runs outside `InspectorClient`, so nothing else puts a proxy
under it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UL9retfvAXRgvi6EWk4SY4
Signed-off-by: cliffhall <cliff@futurescale.com>
- `revokeStoredKeys`' pre-read of `getTokens` sat outside the best-effort
catch. `getTokens` parses through `OAuthTokensSchema`, so a persisted token
that no longer validates rejected and abandoned the local delete `--relogin`
promises — over a grant that could not have been revoked anyway. Caught and
retained as a failed outcome; the clear proceeds.
- The multi-key bookkeeping was wrong twice. A token is now marked spent only
on a `revoked` outcome, so a failed or unsupported attempt no longer skips a
duplicate entry under the other key that may hold the credentials or metadata
that would have worked. And a failure now outranks an earlier success for
reporting, so a stale grant left live cannot be hidden behind the first key's
200.
- `--no-revoke` is now rejected without `--relogin`, ahead of the short-circuit
returns, matching the parser's existing stance on accepted-but-inert flags
(`--strict`). On its own it reads as "this run will not revoke anything",
true only because nothing was being cleared.
- Extracted the App draft merge as `utils/serverWithDraftSettings`, so the rule
has a test. The hook tests receive whatever they are handed and cannot tell a
draft from a persisted entry, and `App.test.tsx`'s settings harness connects
a stdio server where the OAuth section never renders.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
- RFC 8414 §2 defaults an omitted `revocation_endpoint_auth_methods_supported`
to `client_secret_basic`; it does not inherit the token endpoint's list.
Inheriting made metadata advertising only `client_secret_post` there send
POST credentials to an endpoint that never advertised the method.
- The 5s timeout was inert on the web paths: `createRemoteFetch` re-issues the
call as a POST to `/api/fetch` and drops `init.signal`, and the backend's
outbound fetch gets none either, so a wedged AS could hold the teardown open
indefinitely. A wall-clock race now enforces the deadline; the signal is kept
because it genuinely cancels the direct-fetch paths.
- The settings-modal Clear read `settingsModalTarget`, which comes from the
persisted `servers` list, while edits live in `settingsDraft` until the save
debounce. Toggling the checkbox and clearing immediately used the previous
value. It now merges the draft.
- `--relogin` clears both key spellings, so it now revokes from both rather
than stopping at the first populated one — a stale entry under the other
spelling is a live grant. The normalised key goes first, matching
`findStoredServerState`, and a second key holding the same token is skipped.
- The fixture's /oauth/revoke now ENFORCES RFC 7009 §2.1 client
authentication (both RFC 6749 §2.3.1 forms), so the e2e's claim that the
Inspector authenticates correctly is actually tested; two negative cases
assert a wrong secret and an unidentified client are refused.
- Dropped the `as unknown as` casts from the tests: a typed `InspectorLogger`
double, and a spy on the real storage instead of a spread-and-cast.
- Added CLI-level tests through `runCli` for `--no-revoke`, the per-server
`oauth.revokeOnClear`, and the stderr warning, and TUI tests for the
forwarded opt-out and the rendered failure.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
Two review findings.
The stored-token refresh path (`refreshStoredAuthToken`) calls SDK discovery
directly rather than through `InspectorClient.effectiveAuthFetch`, so a state
file holding a refresh token and client information but no `serverMetadata`
still hit the upstream OIDC-schema failure and could not refresh against the
very servers this change is for. Its default `discover` now wraps the global
fetch with the compat shim; a caller-supplied `fetchFn` is left alone. Covered
by a test that injects no `discover` — injecting one would bypass the default
under test — against a real server that 404s the RFC 8414 path and serves plain
RFC 8414 metadata at the appended OIDC path.
`oidcDiscoveryCandidates` matched the well-known path by bare prefix, so it
also claimed `/.well-known/oauth-authorization-server-backup` and would have
replaced that path's failed response with a document fetched from a derived
URL. It now requires the exact path or a `/` boundary, which is precisely the
set `buildDiscoveryUrls` can emit.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UL9retfvAXRgvi6EWk4SY4
Signed-off-by: cliffhall <cliff@futurescale.com>
Clearing the Inspector's OAuth state deleted the local copy of the tokens
and stopped there. The access token — and the refresh token, which is
long-lived by design — stayed valid at the authorization server until they
expired on their own, so a day of connect/disconnect iteration left it
holding grants for sessions that ended hours ago. RFC 7009 §1 describes
exactly that case, and nothing in the repo constructed a revocation
request, so a server author implementing `revocation_endpoint` had no way
to exercise it with the tool that exists to exercise this.
core/auth/revocation.ts builds and sends the request; the three clear paths
call it, so the behavior is written once — web's "Clear OAuth state and
disconnect" (both the active and stored-only paths), the TUI's Auth tab,
and the CLI's --relogin. It runs BEFORE the local clear, because the token,
the client credentials and the discovered `revocation_endpoint` all live in
the store the clear is about to empty. The request names the refresh token
when there is one: §2.1 asks the AS to also invalidate the access tokens
issued under the same grant, so one request covers both halves.
Best-effort by construction. Every path returns a TokenRevocationOutcome
rather than throwing: no advertised endpoint means nothing is sent and that
authorization server behaves exactly as before, and a network error, a
non-2xx, a 5s timeout or an unreadable store is reported and nothing more.
Forgetting the tokens is what the user asked for. `lost_authorization_state`
recovery deliberately skips it — it clears a half-finished flow to retry it,
and an authorization that never completed has no grant to revoke.
Client authentication resolves the preregistered slot first, mirroring
BaseOAuthClientProvider.clientInformation; reading only the dynamic slot
would send no authentication for a server configured with `oauth.clientId`.
The Basic credential is byte-identical to the SDK's applyBasicAuth so this
request and the token request cannot present the same secret differently.
Opt out per server with `oauth.revokeOnClear` (default on, only `false`
written to disk) or per CLI run with --no-revoke. Read at clear time rather
than connect time, so toggling needs no reconnect. Turning it off is a
testing affordance: a client that walks away still holding live tokens is a
case a server author may want to reproduce.
test-servers now advertises `revocation_endpoint` and serves POST
/oauth/revoke, faithful on §2.2 (an unknown token is a success) and §2.1
(revoking a refresh token invalidates the access tokens under the same
grant), with oauth-revocation-http.json / oauth-no-revocation-http.json as
showcase configs.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
The root `ci` script was never what runs in GitHub CI, and the name worked
against us three ways: it collides with the built-in `npm ci` (which does
something else entirely and is one keystroke away), it invites the
cross-engine passes into a workflow by looking like the thing a workflow
should call, and it blurs "GitHub CI" and "the local gate" in every doc
that names it.
- Rename `ci` → `local:gate` and `ci:storybook` → `local:storybook`, with
no back-compat alias: `npm run ci` now fails with npm's missing-script
error, which prints the available scripts.
- Add `scripts/lib/workflow-gate.mjs` + its `test:scripts` suite, which
runs over the repo's real `.github/workflows/**` and fails on an
invocation of a `local:*` script, of a non-Chromium engine pass (derived
from SUPPORTED_BROWSERS), or of `smoke:web:engine`, or on `SMOKE_BROWSER`
set to anything but a literal `chromium`. It does not forbid
`npm run smoke`, `smoke:web:chromium` or `smoke:tui` — those belong in CI
and are there today. Comments are not scanned, so the workflow can still
explain itself. This is the durable half; the rest is prose.
- Document the two-tier model once, in the root README ("Two tiers: GitHub
CI and the local gate"), and point the other docs at it rather than
restating the split.
- Sweep the ~40 stale references across README, AGENTS.md,
.github/copilot-instructions.md, the client READMEs, the roadmap doc and
the script headers. AGENTS.md and copilot-instructions.md change
together, per the mirror rule.
No gate changes what it actually runs.
Signed-off-by: Cliff Hall <cliff@futurescale.com>
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 1.
`refreshStoredAuthToken` consumes the same fallback and made a single SDK
discovery call, so preserving the path there — without the walk the other two
consumers got — would have regressed exactly the case the walk exists to keep:
a path-hosted MCP server whose authorization metadata is published at the
domain root.
It injects its own discovery function as a test seam, so the walk is now also
exposed as `discoverAuthorizationServerMetadataFromCandidates(candidates,
discover)`, which takes that function and reports *which* candidate answered —
the CLI needs the winning URL, since it is the base the token request is made
against, not just the metadata.
Also drops the `as unknown as typeof fetch` from the new test in favour of
`vi.fn<typeof fetch>()`.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013hzwzS8UvBsr4yZm7o7sev
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 8 (its one suppressed finding; no new inline comments).
The `boolean-schema` suggestion for a bare `true` said the replacement
"narrows the schema deliberately". That holds in an ordinary property
position and inverts under a negating parent: `{"not": true}` rejects
every value, so swapping in `{"type": "object", …}` leaves the parent
rejecting only objects and every non-object then passes — the enclosing
contract is widened. Under `if` it redirects which branch applies rather
than tightening anything.
- Parent-aware advice under `not`: suggest `{}`, which is the exact
object form of `true` and is already exempt from `untyped-schema` in
that position, with the message saying why a concrete type would be
wrong there.
- Everywhere else, and in the `untyped-schema` suggestion too, describe
the replacement as a deliberate CHANGE of contract rather than a
narrowing. A suggestion has no business claiming an effect it cannot
guarantee across the positions it can print from.
Tests pin the `not`-specific advice and assert no bare-`true` suggestion
in any of the three positions contains "narrow". The CLI README's sample
report is re-captured from a real run.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UL9retfvAXRgvi6EWk4SY4
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 7.
- Add `contentSchema` to the traversal. It holds the schema the decoded
string content must satisfy, so a bare `true` or a remote `$ref` under
it is a real finding all three clients were walking past. Also admit
the three content keywords to `CONSTRAINING_KEYWORDS` on the same
documented grounds as `format` — annotation-only by vocabulary, but
plainly an attempt to constrain, so reporting such a schema as
accepting anything would be noise. The comment now names that
category rather than singling `format` out.
- Make `handleError` async and await the envelope write. Round 6 made
the `--strict` report await its write and stopped one line short: the
envelope emitted for the very same throw still raced `process.exit()`,
so the machine-readable line a CI caller parses could be lost on a
pipe. The launcher's comment claiming the envelope was "well inside
the pipe buffer" was wrong reasoning and is replaced — fitting in the
buffer says nothing about whether the write was performed. Its call
site now awaits; the bin's `.catch(handleError)` needed no change.
- The CLI README called its sample block the "complete output" when the
error envelope follows it on every non-zero exit. Now scoped to "the
complete report", with the envelope named and linked.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UL9retfvAXRgvi6EWk4SY4
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 6.
- Share one traversal between `walk` and `declaresAnyId`. The `$id` guard
added last round descended through EVERY object value, including
instance data — `examples: [{"$id": "x"}]`, a `default`, a `const` —
where `$id` is a payload key declaring no embedded resource. Reading
one there suppressed `remote-ref` for the whole document, turning last
round's false-positive fix into a blanket false negative. Both passes
now go through `forEachSubschema`, driven by the same keyword tables,
so they cannot drift about what counts as a schema position.
- Await the stderr writes. `writeSchemaLintReport` used a bare
`process.stderr.write`, but both CLI exit paths call `process.exit()`
as soon as the work returns, which discards anything still buffered on
a piped stderr — truncating or losing the multi-block `--strict`
report on exactly the redirected-output runs a CI caller uses. Added
`awaitableError` beside the existing `awaitableLog` and awaited it.
The test's fake defers its callback so a synchronous one cannot make
the assertion vacuous; two existing stderr stubs that never invoked
their callbacks were fixed in the same pass.
- Show the CLI README's sample report in full. It printed one error
block immediately followed by the summary, a shape the formatter
cannot produce — every finding gets a block. Captured from a real run.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UL9retfvAXRgvi6EWk4SY4
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 5.
- The sidebar tooltip applied `hasError` to the TOTAL, so a tool with one
error and three warnings announced as "4 schema portability errors" —
a defect the previous round introduced while adding that wording, and
worse than the vague text it replaced. Both the tooltip and the
accessible label now carry the breakdown, formatted by a new shared
`summarizeToolFindings` so the fourth place to phrase a severity count
cannot drift from the other three.
- Suppress `remote-ref` when the document declares an `$id` anywhere.
`$id` establishes a base URI, so with a root `$id` of
`https://example.com/root` the ref `https://example.com/root#/$defs/x`
resolves to THIS document and needs no fetch — the finding and its
"inline the referenced schema" remediation would both be wrong.
Classifying correctly needs full RFC 3986 resolution against nested
embedded resources, which this module does not do, so it declines
rather than guessing. A missed finding is the acceptable direction.
- Run the `--strict` validations before every short-circuit return.
`servers/list`, `servers/show`, `--list-stored-auth` and
`--print-handoff` all return from `parseArgs` before any connect, so
the checks further down let `--strict --method servers/list` succeed
while silently ignoring a tools/list-only flag.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UL9retfvAXRgvi6EWk4SY4
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 4.
- Add `SCHEMA_UNPORTABLE` to `codeForExit`. `CliExitCodeError` lets a
caller omit the envelope, so without the case an envelope-less throw
reported the generic `error` — the same silent mislabelling the rename
was meant to prevent, one layer down. Covered via the envelope-less
construction that exposed it.
- Keep a local `type` (and the preset's `required`) alongside the remote
`$ref` in the showcase config. The TUI's `schemaToForm` dispatches on
`type`, so a `$ref`-only property fell through to a string field that
`add`'s numeric handler then rejected — making the README's "every
showcase tool is runnable" claim false there. The `remote-ref` finding
is unaffected, since that rule reads `$ref` independently.
- Restrict `type-union` to well-formed unions: non-empty, unique, and
all recognized JSON Schema type names. Its message asserts the
construct is *legal* JSON Schema, which is false for `[]`, `[3]` or
`["bananas"]`, and the `anyOf` it suggested for those would be invalid
too. Malformed schemas stay the SDK parser's business.
- Drop `if`/`then`/`else` from the presence-only allowlist. They only
constrain as a pair, so `{"if": {"const": 1}}` accepts every value and
was evading `untyped-schema`; the pairing is now checked explicitly.
- Distinguish the web sidebar severities by SHAPE, not just colour — a
circle for an error, a triangle for a warning — and name the severity
in the tooltip. Colour alone left a colour-blind reader unable to tell
them apart.
- Fix the stale "across 2 tools" count in the CLI README's hint sample.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UL9retfvAXRgvi6EWk4SY4
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 3.
- Rename the `--strict` failure from "invalid" to "unportable"
(`EXIT_CODES.SCHEMA_UNPORTABLE`, envelope `schema_unportable`, and the
message text). The module argues at length that these schemas ARE
valid JSON Schema and are merely refused by some clients, so a
machine-readable code of `schema_invalid` told automated callers the
opposite. Done now, while nothing depends on it.
- Add a fixture-path integration case. Every other test built a
`ServerConfig` by hand, so `resolveConfig`'s `rawToolSchemas`
forwarding was uncovered and `unportable-schemas-http.json` was loaded
by nothing — the shipped fixture could have lost its overrides with
the suite still green. It now goes through loadConfig/resolveConfig
and asserts the exact per-tool verdict the README claims.
- Refresh the CLI README's sample output, which still showed the
pre-round-2 wording and fixture, and so dropped the contract warning
that rewording existed to add. Captured from a real run rather than
written by hand, plus a paragraph making the suggestion pattern
explicit.
- Drop the self-referential "[Try it](#unportable-tool-schemas)" link in
the root README, which pointed back at its own section.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UL9retfvAXRgvi6EWk4SY4
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 2.
- Remove `non-object-root` entirely. It could never fire: the SDK types
`inputSchema` with `type: literal("object")`, so a tool with any other
root fails `ListToolsResultSchema` and `salvageListItems` drops it
before any client sees it. Verified against a live server advertising
`inputSchema: {type: "array"}` — `listAllTools()` returns []. A rule
that cannot fire is worse than none, because the docs then claim a
check the tool does not perform; the condition is already reported
through the malformed-items surface. Pinned by an integration test so
a change in that behavior is a signal rather than a silent gap.
- Fix the bare-`false` suggestion. `properties: {a: false}` forbids the
property; deleting the entry *permits* it with any value under the
default `additionalProperties`, so the advice inverted the schema it
replaced. Now recommends `{"not": {}}` and says why deletion differs.
- Exempt `not`-parented schemas from `untyped-schema`, so the `{"not":
{}}` above does not trip this module's own rule — under negation an
always-accepting subschema means always-reject, making "accepts any
value" the exact opposite of what it does. Scoped to that position,
not inherited by its subtree.
- Say plainly, in the `true` and untyped suggestions, that they narrow
the schema deliberately rather than implying an equivalent rewrite.
`true` has no portable object-form equivalent — `{}` is the very shape
this lint flags.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UL9retfvAXRgvi6EWk4SY4
Signed-off-by: cliffhall <cliff@futurescale.com>
- Scope `non-object-root` to `inputSchema`. The pinned SDK types
`inputSchema` with `type: literal("object")` but `outputSchema` as a
bare `looseObject`, so a `{"type":"string"}` output is conforming; the
rule was claiming a requirement MCP does not make, at error severity.
- Decide `untyped-schema` from an allowlist of constraining keywords
rather than a denylist of annotations. JSON Schema ignores keywords it
does not recognize, so `{"vendorHint": true}` accepts every value and a
denylist passed it as constrained. `$defs`-only schemas now flag too.
- Walk `dependentSchemas` and draft-07 `dependencies`, with a test
pinning that the property-name-array form is not a schema position.
- Suggest `anyOf` branches for an array-form `type`, never "drop it from
`required`" — accepting `null` and permitting absence are independent
contracts, so the old suggestion silently changed the schema's meaning.
- Reject `--strict` alongside `--app-info`. That path returns NDJSON from
`runMethod` and never reaches the lint, so the flag was a no-op gate
that could never fail.
- Colour the web sidebar flag from the `--inspector-*` severity tokens,
matching the detail panel and avoiding the default yellow that failed
the a11y check.
- Add a live integration test for `rawToolSchemas`. It caught that an
`outputSchema` override on a preset returning no structured content
makes every call to that tool fail client-side validation — the
showcase config did exactly that, so the flagged tool could not be
run. Moved the bare `true` onto `get_temp`, documented the caveat, and
pinned both halves.
- Drop the double cast from the lint test fixture.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UL9retfvAXRgvi6EWk4SY4
Signed-off-by: cliffhall <cliff@futurescale.com>
Closes#1005
A tool schema can be perfectly legal JSON Schema and still be rejected by
the client the server is meant to run against — Go's `jsonschema` emits a
bare `true` for `interface{}`, which the SDK accepts and Claude Code
refuses with an opaque "Invalid input". The Inspector passed it through
silently.
Adds `core/json/schemaLint.ts`, a portability lint over a tool's
inputSchema/outputSchema, and surfaces one shared verdict everywhere:
the CLI's `--strict` report (exit 6 on an error-severity finding), the
TUI's tool list marker + detail pane, and the web Tools sidebar flag +
detail section.
Deliberately not a JSON Schema validator: a census of 617 public servers
found zero that fail the SDK's own parse, so a conformance check would
report nothing on real servers. Each rule is instead a construct that is
legal JSON Schema and known to be refused or mishandled by a shipping
client — a bare boolean in a schema position, an array-form `type`, a
remote `$ref`, a schema with no validation keyword, a non-object root.
`test-servers/configs/unportable-schemas-http.json` exercises every rule,
via a new `rawToolSchemas` override — the Zod-built presets cannot emit
any of these shapes, which is the same reason a real server only hits
this when its schemas come from another generator.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UL9retfvAXRgvi6EWk4SY4
Signed-off-by: cliffhall <cliff@futurescale.com>
`npm audit` (dev included) reported a low-severity advisory against
`esbuild` in three of the four client installs. It is not in the
production tree — `npm audit --omit=dev` has been clean everywhere since
2058 — but a permanently non-clean audit trains people to skim the
channel, which is the same argument 2058 made.
`tsup@8.5.1` declares `esbuild: ^0.27.0`, and the advisory covers
0.27.3 - 0.28.0 with 0.27.7 the last 0.27.x, so there is no upward
escape inside tsup's declared range. `npm audit fix` "resolves" it by
silently downgrading to 0.27.2 across three installs — the trap hit and
reverted during 2058. An `overrides` entry forces the copy above the
range instead.
- Add `"overrides": { "esbuild": "^0.28.2" }` to clients/{web,cli,tui}.
That also covers the tui's `tsx` copy (0.28.0), so no separate update
is needed.
- Re-resolve each lockfile. Every changed entry is an esbuild one: the
nested `tsup/node_modules/esbuild` and `tsx/node_modules/esbuild`
copies collapse into the single deduped 0.28.2, and no other package
moves.
- Document the pin in AGENTS.md, including when to drop it.
The real gate here is the build, not the audit: this puts tsup on an
esbuild major it does not declare. `npm run build` succeeds for all
three clients, and `npm run ci` passes.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
The web client and the TUI both read `savedSettings.metadata` and pass it
as `InspectorClientOptions.defaultMetadata`. The CLI never did, and
`InspectorClient` only reads the option rather than falling back to
`serverSettings.metadata` — so a per-server `metadata` entry was silently
inert for `--cli` runs.
Pass it beside the existing `roots:`, guarded on a non-empty map the same
way web and the TUI guard theirs, so `{}` still means "no defaults".
Precedence is unchanged: call-time keys override defaults in `mergeMeta`,
so `--metadata` stays per-invocation and wins on a key collision.
Covered by three catalog-driven tests asserting against the test server's
recorded wire requests. `createTestConfig` is retyped `StoredMCPServer` —
the accurate on-disk shape — since a fixture that cannot express the
per-server keys cannot cover the settings that reach the client.
Closes#2093
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
Copilot review round 2 on #2100, plus a collision the review did not find.
The default app-origin port was 6276 — which is already
RUNNER_OAUTH_CALLBACK_DEFAULT_PORT, the CLI/TUI loopback OAuth callback.
That port is fixed precisely so an OAuth app can pre-register
http://127.0.0.1:6276/oauth/callback, which makes it the one listener here
that cannot move when something takes it first. A running --web would hold
it and break a later --cli/--tui OAuth flow at its registered redirect URI,
and docs/mcp-app-review.md has those two running side by side. This
listener has a documented dynamic fallback, so this listener yields: the
default is now 6278 (6277 skipped — v1's retired proxy port, which the
migration guide tells people to stop forwarding). A test asserts the two
defaults stay distinct.
Review comments:
- POST /api/app-document now applies hono's bodyLimit BEFORE parsing. The
8MiB check ran after c.req.json() had buffered and parsed the whole
request, and bounded only `html` — so bulk in `csp`, or in keys the route
never reads, was fully materialized before the 413.
- `csp` is validated as an HTTP header value and length-bounded. The
controller hands it to res.writeHead() verbatim, where Node throws
ERR_INVALID_CHAR synchronously — inside the handler for a later,
unrelated request, outside this route's error handling.
- The app-origin port now propagates everywhere the other two do: the CLI
--print-handoff port-forward command (and its test), the Dockerfile's ENV
and EXPOSE, and the remote-review SSH recipe in docs/mcp-app-review.md.
Without it an app declaring _meta.ui.domain is unreachable through every
documented remote workflow.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0177mPHAdECD18nLLCTwR5rh
Signed-off-by: cliffhall <cliff@futurescale.com>
One conflict, in `clients/tui/__tests__/App.test.tsx`: both sides added an
`opts` capture to the test harness's `FakeClient` for different fields —
`defaultMetadata` on v2/main (#1910), `environment.fetch` here (#2067) — so the
resolution is the union of the two, keeping both rationales and both tests.
Everything else auto-merged. `npm run ci` passes on the result.
Signed-off-by: cliffhall <cliff@futurescale.com>
Addresses review round 7 on #2099 (three suppressed comments).
- Proxy variables are now cleared for the WHOLE of `transport.test.ts` and
`cli.test.ts`, not just around the proxy tests. Measured: undici's
`EnvHttpProxyAgent` captures its proxy URLs when it is CONSTRUCTED and re-reads
only NO_PROXY per request, so with the process-wide memo, any earlier test that
connects on a machine with an ambient proxy would pin the agent to it and a
later test setting HTTP_PROXY would be talking to an agent that is not
listening. Clearing up front means nothing can build the singleton before the
proxy tests do, and restoring only in the file-level afterAll means the memo
cannot be left bound to a closed test proxy.
Both suites verified green under a hostile ambient env
(`HTTP_PROXY`/`http_proxy`/`HTTPS_PROXY` pointed at a dead port). In fairness
this is hardening rather than a fixed failure: the concrete cross-test break
did not reproduce with the current test order — but it depends on test order
and on the developer's environment, which is exactly what this removes.
- The guard's failure message said a package "is declared external but was
inlined" even for a candidate that came only from the root dependencies —
i.e. #2067's own shape, where the missing `external` entry IS the bug, so the
message contradicted the config it was asking the reader to check. It now
words the failure by candidate source, and says to add the entry.
The capture-at-construction behavior is documented on `createProxyFetch`, since
it is the reason the suites are structured this way.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JTHVxSu8AUgHRLvo1ntZ8H
Signed-off-by: cliffhall <cliff@futurescale.com>
Addresses review round 5 on #2099 (three suppressed comments, no posted ones).
The `fetch: createProxyFetch()` line in each Node client is the whole of that
client's proxy support — it is what puts the proxy at the BOTTOM of the fetch
stack, where InspectorClient's wrappers compose over it rather than discarding
it. `InspectorClient` sets `fetchFn` unconditionally, so the transport-level
fallback never fires for these clients. Both lines could have been deleted with
every existing test still green.
- CLI: a test that runs a real `tools/list` against a real HTTP test server with
HTTP_PROXY pointed at a real forwarding proxy, asserting the request traversed
it. Behavioral rather than structural on purpose — the defect this replaces was
a runtime dispatch failure no structural assertion would have caught. It clears
all six proxy vars first so an ambient value cannot route it elsewhere.
- TUI: the mock added earlier always returned undefined, so the proxy branch was
never exercised. `createProxyFetch`'s return is now controllable, and two tests
assert the sentinel reaches `environment.fetch` when a proxy is configured and
that the slot stays unset when one is not.
Both were confirmed to fail with their client's wiring line removed.
Also updates the `InspectorClientEnvironment.fetch` doc in core/mcp/types.ts,
which still told callers that Node leaves this undefined and uses the global
fetch.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JTHVxSu8AUgHRLvo1ntZ8H
Signed-off-by: cliffhall <cliff@futurescale.com>
Addresses review round 4 on #2099 — three posted comments and three suppressed
ones, all valid.
- A rebuilt Response reported `url: ""`, `redirected: false`, and
`type: "default"` even after undici followed a redirect: a fetch-contract
regression visible only to proxy users, and one `core/auth/endpointOverrides.ts`
explicitly guards against by declining to rebuild at all. This adapter has no
such choice — the response arrives from the wrong realm — so it now reinstates
the three with `defineProperty`.
- Normalizing a `Request` input dropped every option the Request itself carried
beyond method/headers/body/signal. Most visibly `redirect`: a
`new Request(url, { redirect: "manual" })` started following redirects instead
of surfacing them. All of them are now carried through, with an explicit
`init` still winning per the fetch contract.
Both are covered by tests driving a real redirect through a real proxy, and
both were confirmed to fail without the fix.
- `@modelcontextprotocol/ext-apps` was added only to the CLI's `external` list.
AGENTS.md requires every root-declared runtime dependency in all three, since
which client reaches it is a function of what `core/` imports rather than of
the client's own code. Added to the web runner and TUI lists too.
- The `BUNDLED_CLIENTS` test compared the hard-coded list against a hard-coded
copy of itself, so a new tsup-bundled client would have left it green while
going unchecked. It now discovers clients from disk (a client is tsup-bundled
iff it ships a tsup config) and compares. Verified it fails when a stub client
is added.
- AGENTS.md's project tree still described `core/mcp/node/` as only the stdio
transport factory; it now carries `proxyFetch.ts` and its role.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JTHVxSu8AUgHRLvo1ntZ8H
Signed-off-by: cliffhall <cliff@futurescale.com>
Addresses the suppressed review comment on #2099, which was the sharpest of the
three rounds: the guard stopped looking at a package the moment its `external`
entry was removed — so deleting `"undici"` from the web config would have made
the pre-fix bundle pass. That is self-referential, and #2067 was precisely a
MISSING declaration, so the guard covered its regression but not the bug.
Candidates are now the union of two independently-derived lists: each client's
`external` array, and the root manifest's `dependencies` — root-declared by
rule, reached through `core/`, and required to be external in every client
bundle. Removing a package from an `external` list can no longer remove it from
scrutiny.
That immediately surfaced one latent instance of the same rule violation:
`@modelcontextprotocol/ext-apps` is root-declared and reached through
`core/mcp/apps.ts`, but was absent from the CLI's `external` list and was being
inlined into `index.js` — dragging the transitive v1 SDK and zod-to-json-schema
in with it. It is ESM, so it was not failing the way `undici` did; what it
violated was the rule. Adding it drops the CLI bundle from 590KB to 518KB.
It is also a real example of the shape the previous revision could not see: it
was inlined into the entry, emitting no `<pkg>-HASH.js` chunk at all.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JTHVxSu8AUgHRLvo1ntZ8H
Signed-off-by: cliffhall <cliff@futurescale.com>
