build(deps): bump the go-modules group with 2 updates

[dependabot]

Bumps the go-modules group with 2 updates: github.com/modelcontextprotocol/go-sdk and modernc.org/sqlite.

Updates github.com/modelcontextprotocol/go-sdk from 1.7.0 to 1.8.0

Release notes

Sourced from github.com/modelcontextprotocol/go-sdk's releases.

v1.8.0

This release is equivalent to v1.8.0-pre.2. Thank you to those who tested the pre-release.

In this release we introduce several fixes and improvements on top of v1.7.0. It adds no new protocol revision: the supported set is unchanged, and 2026-07-28 remains the newest version the SDK negotiates.

The bulk of the work is hardening the transports against resource exhaustion, closing session leaks, deadlocks and teardown hangs found by users running the new protocol at scale, and giving servers explicit control over which protocol versions they advertise.

Two behavior changes are guarded by new MCPGODEBUG flags; see the section below.

Hardening against resource exhaustion

Every decoding path that buffers incoming input is now bounded. JSON payloads are rejected past 1000 levels of nesting, before the parser recurses. Both SSE readers cap the bytes buffered for a single event via MaxEventSize on SSEClientTransport and StreamableClientTransport, and the stdio transport caps a single JSON-RPC frame via StdioTransport.MaxLineLength.

On the OAuth side, dynamic client registration responses are bounded to 1 MB, and the discovery code now validates metadata documents rather than trusting them.

• internal/json: json nesting depth check by @​yarolegovich (#1200)
• mcp: add an sse event size cap by @​yarolegovich (#1205)
• mcp: add a max request body size of the old transport by @​yarolegovich (#1224)
• oauthex: bound dcr read to 1mb by @​yarolegovich (#1206)
• oauthex: oauth discovery checks by @​yarolegovich (#1220)

Restricting the protocol versions a server supports

ServerOptions.SupportedProtocolVersions lets a server narrow the set of versions it advertises and negotiates. The list can only narrow, never widen; naming a version the SDK does not implement panics at construction. Relatedly, a stateful streamable handler receiving a 2026-07-28 request now returns that same JSON-RPC error instead of a plain-text 400, so the client can renegotiate down instead of losing the connection.

• mcp: allow restricting the protocol versions a server supports by @​guglielmo-san (#1199)
• mcp: return json-rpc error on stateful server serving protocolVersion20260728 requests by @​guglielmo-san (#1143)
• mcp: limit SSE discover supportedVersions to SSE-defined revisions by @​arimu1 (#1121)
• mcp: prefer the negotiated protocol version over the request context by @​gwyszynski-sa (#1163)
• mcp: export ClientSessionOptions.ProtocolVersion by @​arimu1 (#1127)

Per-request cache control

ServerOptions.SetCacheable is a new hook that decides the ttlMs and cacheScope fields of every result carrying them: server/discover, the four list methods, and resources/read. It runs once per result, after the handler returns, with the values that handler produced, so it can set a policy globally while still letting an individual handler override it. Anything left unset falls back to the protocol default of public.

• mcp: allow per request Cacheable customization by @​guglielmo-san (#1203)

Behavior changes guarded by MCPGODEBUG

Two new escape-hatch flags restore the previous behavior of the changes above. Both will be removed in v1.9.0.

• plaintextstatefulrejection=1 — restore the plain-text http.Error 400 body a stateful StreamableHTTPHandler previously returned for a request carrying per-request metadata. The default is now a JSON-RPC -32022 CodeUnsupportedProtocolVersion error with an UnsupportedProtocolVersionData payload advertising the legacy versions the server supports. Introduced by #1143.
• blockingcancelnotify=1 — restore the previous behavior where a cancelled call waits synchronously for notifications/cancelled to be delivered (up to 5s) before returning, joining any delivery error into the caller's error. The default now retires the call immediately and sends the notification asynchronously. Introduced by #1151.

Options below were removed, according to plan:

• seterroroverwrite
• enableoriginverification
• disablecontenttypecheck

... (truncated)

Commits

• 3f3b699 refactor: remove legacy MCPGODEBUG compatibility for new protocol release (#1...
• 830f0b7 mcp: update conformance tests (#1231)
• 12cbafe oauthex: oauth discovery checks (#1220)
• 3632967 mcp: add an sse event size cap (#1205)
• 0d3036f mcp: allow per request Cacheable customization (#1203)
• cb0de64 mcp: add a max request body size of the old transport (#1224)
• 2fdabde mcp: do not check metatada on notifications (#1215)
• 59185e6 build(deps): bump ossf/scorecard-action from 2.4.3 to 2.4.4 (#1217)
• a6764cf build(deps): bump golang/govulncheck-action from 1.0.4 to 1.1.0 (#1218)
• 8227246 fix: only subscribe when server advertises capability (#1221)
• Additional commits viewable in compare view

Updates modernc.org/sqlite from 1.58.0 to 1.59.0

Changelog

Sourced from modernc.org/sqlite's changelog.

Changelog

Entries for v1.38.1 through v1.44.1 and for v1.49.1 were added on 2026-09-05, reconstructed from the git history and the merge requests they cite; they were missing at release time.

• 2026-09-15 v1.59.1:

  • Document SQLite's own URI query parameters on Driver.Open: mode, cache, immutable, nolock, psow and modeof. Every connection is opened with SQLITE_OPEN_URI, so in a DSN starting with file: these have always worked; only the driver's own keys were listed. The docstring also spells out the trap that a plain file name has its query stripped before SQLite sees it, so /path/to.db?mode=ro opens read-write. Documentation only.
  • Resolves [GitLab issue #257](https://gitlab.com/cznic/sqlite/-/issues/257).
  • Ship a Software Bill of Materials: sbom.cdx.json (CycloneDX 1.6) and sbom.spdx.json (SPDX 2.3), both validated against the published schemas, with SBOM.md explaining what they cover. They name what an SBOM built from the module graph cannot see -- the transpiled SQLite 3.53.4 and sqlite-vec C, and the upstreams modernc.org/libc vendors, musl among them -- and mark every component as linked into your binary, test-only, or compiled into nothing. Documentation only.
  • Ship LICENSE-3RD-PARTY.md, a transitively flattened inventory of every third-party component this module carries: the whole Go module graph, the transpiled SQLite and sqlite-vec C that no go.mod names, and the upstreams modernc.org/libc carries in turn, musl among them. It reproduces all seventeen distinct license texts in full and separates what is linked into your binary from what only appears in the module graph. The LICENSE name prefix is what makes go mod vendor carry it into downstream vendor/ trees. Documentation only.

• 2026-09-15 v1.59.0:

  • Bump the pinned modernc.org/libc to v1.75.7 and re-vendor lib/ and vec/. The transpiled SQLite is unchanged, still 3.53.4. On the Linux targets the new libc replaces transpiled musl memcpy, memmove, memset, memcmp and strlen with native Go, cutting CPU time on query-heavy workloads by up to a third; see the new Performance section below. As always, downstream go.mod files must pin the same modernc.org/libc version this repository's go.mod does; see [GitLab issue #177](https://gitlab.com/cznic/sqlite/-/issues/177).
  • Hand user-defined function and aggregate callbacks a pooled *FunctionContext instead of allocating a fresh one per call, removing the last driver-side allocation per invocation. Like the argument slice, it is valid only for the duration of the callback and must not be retained past its return. Updates [GitLab issue #226](https://gitlab.com/cznic/sqlite/-/issues/226). See [GitLab merge request #137](https://gitlab.com/cznic/sqlite/-/merge_requests/137).
  • Add regression tests pinning the identity and the pooling of that context. See [GitLab merge request #138](https://gitlab.com/cznic/sqlite/-/merge_requests/138), thanks Ian Chechin!
  • Add a Performance section to the package documentation: measured CPU-time ratios of this driver against the same SQLite compiled from C, where the gap comes from, and the two consequences for applications — index the columns that ORDER BY, GROUP BY and WHERE use, and bound the database/sql pool with SetMaxOpenConns.

• 2026-09-01 v1.58.0:

  • Upgrade to SQLite 3.53.4. It carries upstream's own fix for the journal-rollback data-corruption bug, so the local super-journal patch v1.56.0 introduced is dropped; recovery behavior is unchanged. Also bumps the pinned modernc.org/libc to v1.75.6; as always, downstream modules must pin the same version this one does, see [GitLab issue #177](https://gitlab.com/cznic/sqlite/-/issues/177).
  • Add opt-in support for Linux Open File Description (OFD) locks on database files, off by default; without opting in, locking behavior is byte-for-byte that of previous releases. A POSIX record lock is owned by the (process, inode) pair, so any Close of any descriptor of the database file anywhere in the process silently strips SQLite's locks; OFD locks survive that. Enable it process-wide with MODERNC_SQLITE_OFD_LOCK=1 in the environment, or with the new OFDLocking(true) before the first connection is opened; OFDLockingEnabled reports the mode in effect, and the new ErrOFDLockingTooLate and ErrOFDLockingUnavailable report a switch attempted too late and a platform or filesystem without the feature. Why it is process-wide rather than a DSN parameter, what WAL's -shm coordination still uses, and the /proc/locks measurements behind the design are in [GitLab issue #255](https://gitlab.com/cznic/sqlite/-/issues/255).
  • Resolves [GitLab issue #255](https://gitlab.com/cznic/sqlite/-/issues/255). See [GitLab merge request #136](https://gitlab.com/cznic/sqlite/-/merge_requests/136), thanks Nathan Herring (@​technosloth), and thanks Gani Georgiev (@​ganigeorgiev) for pressing the opt-in default!

• 2026-08-19 v1.57.0:

  • Add an opt-in _defensive DSN query parameter turning on SQLite's defensive mode for the connection. On such a connection PRAGMA writable_schema=ON, PRAGMA journal_mode=OFF and PRAGMA schema_version=N become silent no-ops, and writes to a virtual table's shadow tables and to sqlite_dbpage fail. It is a hardening measure, not a sandbox for hostile database files, for which it is only one of the steps SQLite recommends, and it is a property of the connection, not of the file. Absent, or _defensive=0, nothing changes.
  • Reject _defensive=1 together with _journal_mode=OFF (or _journal=OFF) instead of opening a connection in which neither was honoured: SQLite turns that PRAGMA into a no-op that still reports success. Only DSNs using the new parameter can be affected. See [GitHub pull request #6](modernc-org/sqlite#6), thanks wsman!
  • Ship the sqlite-vec license notice this module has been missing since vec/ arrived in v1.47.0. sqlite-vec is Copyright (c) 2024 Alex Garcia, dual-licensed Apache-2.0 OR MIT and used here under MIT; the text now ships as LICENSE-SQLITE_VEC, and make vendor fails rather than quietly dropping it.
  • The SQLite notice is renamed from SQLITE-LICENSE to LICENSE-SQLITE; update any direct links to it. Its contents are unchanged. The rename is what makes go mod vendor carry both notices into downstream vendor/ trees: it selects license files by name prefix, so a name merely ending in LICENSE was never propagated.
  • Let a caller-constructed Driver register its own functions, collations and virtual table modules, through new RegisterFunction, RegisterScalarFunction, RegisterDeterministicScalarFunction, RegisterCollationUtf8 and RegisterModule methods plus Must* variants, and let vtab.RegisterModule honour its db argument. Behavior change: vtab.RegisterModule(db, ...) where db was opened on a caller-constructed Driver used to discard db and land on the registered sqlite driver, reaching every connection in the process; it now lands on that Driver alone, so a sql.Open("sqlite") connection that used to resolve such a module gets no such module. Everything else is additive, and the isolating change discussed in [GitLab issue #254](https://gitlab.com/cznic/sqlite/-/issues/254) is deliberately not made here. See [GitLab merge request #135](https://gitlab.com/cznic/sqlite/-/merge_requests/135), thanks Ian Chechin!
  • Promote freebsd/386, freebsd/arm and netbsd/amd64 from experimental to fully supported. The package documentation's platform table had carried seventeen entries while this module shipped, cross-built and tested twenty; all three have been in the builder matrix since v1.53.0 and pass the full suite on this release's commit. Documentation only — lib/ is byte-for-byte what v1.56.0 shipped.

• 2026-08-03 v1.56.0:

  • Re-vendor the transpiled sources, picking up modernc.org/libsqlite3's patch for an upstream data-corruption bug in SQLite 3.53.3's journal rollback. A crash during the commit of a multi-database (ATTACH) transaction can leave a hot journal whose zeroed super-journal name still validates, so pager_playback() deletes it without playing it back and leaves the database corrupted. Not a transpilation artifact: a plain gcc build of stock 3.53.3 fails on the same bytes. The SQLite version is unchanged at 3.53.3, every supported target carries the patch, and it will be dropped once upstream ships its own fix.
  • Two targets change beyond that patch. linux/s390x now allocates C bit-fields MSB-first as the big-endian ABI requires, from modernc.org/cc/v4 v4.29.1. linux/riscv64 was regenerated on a host running GCC 11.4.0 rather than 13.3.0, which drops some unreferenced compiler-predefined macro constants and changes what PRAGMA compile_options reports; no SQLite code generation differs. Every other target is byte-identical to v1.55.0 apart from the patch above.
  • Bump the pinned modernc.org/libc to v1.74.4 and the remaining dependencies to their current releases. v1.74.2 and v1.74.3 are retracted upstream over a freeaddrinfo lock leak that deadlocks name resolution, and v1.74.4 is the fix. As always, downstream modules must pin the same modernc.org/libc version this one does, see [GitLab issue #177](https://gitlab.com/cznic/sqlite/-/issues/177).
  • Add NewConnector, returning a database/sql/driver.Connector for use with sql.OpenDB. It opens the same connections sql.Open("sqlite", dsn) does, from the same registered driver, so every function, collation, connection hook and virtual table module registered through this package applies to them. It exists for callers that need to interpose on the physical connections database/sql opens — tracing, metrics, connection-scoped setup — for which the alternative was sql.Register, which is process-global, panics on a repeated name and cannot be undone. A runnable sample is in examples/connector. Resolves [GitLab issue #253](https://gitlab.com/cznic/sqlite/-/issues/253), thanks Alessandro Segala (@​ItalyPaleAle)!
  • Documentation sweep. openbsd/amd64 and openbsd/arm64 join the supported platforms table, the vfs DSN query parameter is documented alongside the others on Driver.Open, the stale go generate and GO_GENERATE instructions are replaced by modernc.org/libsqlite3 and make vendor, and vec and vfs gained the package doc comments they were missing. A caller-constructed sqlite.Driver is now documented as not being the driver this package registers as "sqlite": it carries none of the package-level functions and collations, so it can evaluate upper(x) or date(x) differently. Documentation only.

• 2026-07-20 v1.55.0:

  • Add github.com/mattn/go-sqlite3-compatible shorthand DSN query parameters to ease migration from that driver: _busy_timeout/_timeout, _foreign_keys/_fk, _journal_mode/_journal, _synchronous/_sync, _auto_vacuum/_vacuum and _query_only, each setting the correspondingly named PRAGMA. Values are validated against the same set mattn/go-sqlite3 accepts and an unrecognized one fails the connection, so a typo cannot silently downgrade durability. Review DSNs carried over from that driver before upgrading: a key prior releases ignored entirely now takes effect, so _foreign_keys=on begins enforcing constraints and _journal_mode=wal persistently converts the file; and a duration-style _busy_timeout=5s now fails an open that previously succeeded. See [GitLab merge request #134](https://gitlab.com/cznic/sqlite/-/merge_requests/134), thanks Toni Spets (@​beeper-hifi) and Ian Chechin!
  • Validate every DSN query parameter before applying any of them. Parameters were checked as each was reached, so a DSN rejected for a later parameter had already executed the PRAGMAs ahead of it: file:x.db?_journal_mode=wal&_synchronous=bogus failed the connection and yet left x.db converted to WAL. A failed Open now leaves the database as it found it. The accepted values are unchanged. _pragma remains the sole exception, since its values are executed verbatim and can only be rejected by SQLite as they run.

• 2026-07-15 v1.54.0:

  • Upgrade to SQLite 3.53.3. This also bumps the pinned modernc.org/libc to v1.74.1; as always, downstream modules must pin the exact same modernc.org/libc version this module's go.mod pins (see [GitLab issue #177](https://gitlab.com/cznic/sqlite/-/issues/177)).
  • Under the opt-in _texttotime DSN parameter, best-effort parse date-shaped TEXT values from columns SQLite reports with an empty declared type — aggregates and expressions over a date column (MAX(d), COALESCE(d, ...), upper(d), d || ''), subqueries, and typeless real columns (CREATE TABLE t(x)) — into time.Time, instead of delivering them as a raw string that Scan cannot store into a *time.Time. The existing declared DATE/DATETIME/TIME/TIMESTAMP path is unchanged; this only adds the empty-decltype case. The conversion is strictly best-effort: a value that does not parse as a time falls through to the original string, so no Scan that worked before can newly fail. ColumnTypeScanType continues to report string for empty-decltype columns, since the declared type cannot prove the column is temporal. Without _texttotime the behavior is byte-for-byte unchanged. Resolves [GitLab issue #248](https://gitlab.com/cznic/sqlite/-/issues/248).
  • See [GitLab merge request #133](https://gitlab.com/cznic/sqlite/-/merge_requests/133), thanks Ian Chechin!

• 2026-06-21 v1.53.0:

  • Add experimental netbsd/amd64 support, resolving the long-standing build break in [GitLab issue #246](https://gitlab.com/cznic/sqlite/-/issues/246). This target is intentionally not yet listed among the supported platforms in the package documentation: the port had been broken for years and is only now revived, and there is as yet no real-world experience running it under production workloads. Green CI is not the same as battle-tested — so while the full test suite (including the pcache and vec packages and the -race concurrency test) passes on NetBSD 10.1 / Go 1.26.3, and the entire upstream toolchain (libc, cc, ccgo, libz, libtcl8.6, libsqlite3, libsqlite_vec) is green on the NetBSD CI builder, the target is offered for evaluation only. If you run NetBSD, please exercise it with your own workloads and report back via #246; the intent is to promote it to a fully supported platform after a period of broader real-world testing (on the order of a month) elapses without surprises.
  • Implementation notes: the previously shipped lib/sqlite_netbsd_amd64.go was a stale old-generator transpile that no longer compiled (the mu.enter/mu.leave break in #246); it is replaced by a fresh new-generator transpile consistent with every other platform, and modernc.org/sqlite/vec (sqlite-vec) is vendored and auto-registers on netbsd. Correct operation requires the matching pinned modernc.org/libc, which carries two NetBSD-specific fixes found during this work: the mmap(2) PAD-argument ABI (without it, concurrent WAL access faults with SIGBUS in the WAL-index shared memory) and a working abort(3) (the prior stub left SQLite's crash-recovery writecrash test unable to terminate by signal). As usual, downstream modules must pin the exact modernc.org/libc version this module's go.mod pins.
  • See [GitLab merge request #82](https://gitlab.com/cznic/sqlite/-/merge_requests/82), thanks Leonardo Taccari (@​iamleot) and Thomas Klausner (@wiz)!
  • Add experimental freebsd/386 and freebsd/arm support. As with the netbsd/amd64 target above, these two 32-bit FreeBSD ports are intentionally not yet listed among the supported platforms in the package documentation: freebsd/386 previously shipped a stale, effectively untested SQLite 3.41 transpile, and freebsd/arm is entirely new, so neither has real-world production mileage yet. Both are now freshly transpiled at SQLite 3.53.2 consistent with every other platform, build cleanly, and pass the full test suite (core, WAL/concurrency, and the vec package) on the FreeBSD CI builders; they are offered for evaluation only. If you run 32-bit FreeBSD, please exercise these targets with your own workloads and report back — the intent is to promote freebsd/386, freebsd/arm, and netbsd/amd64 to fully supported platforms in a future release cycle, once a period of broader real-world testing elapses without surprises.
  • Implementation notes: correct operation on freebsd/arm requires the matching pinned modernc.org/libc (v1.73.4), which fixes the per-arch mmap(2) off_t encoding for 32-bit FreeBSD; without it the WAL shared-memory mapping faults with SIGBUS under concurrent access, the same class of bug found on the netbsd port. As usual, downstream modules must pin the exact modernc.org/libc version this module's go.mod pins.

... (truncated)

Commits

• c96a4e6 CHANGELOG.md: document the libc v1.75.7 bump and the Performance section
• 35c446e update dependencies, make vendor
• 9ad02ac Merge branch 'udf-context-tests' into 'master'
• d6d84a9 sqlite: pin the identity and the pooling of the FunctionContext in tests
• 59ec397 doc.go: add a Performance section with measured driver-vs-C ratios
• 51cbcbf Merge branch 'udf-call-pool' into 'master'
• e390a0c sqlite: pool the FunctionContext handed to UDF and aggregate callbacks
• 38af581 doc.go, README.md: point readers at CHANGELOG.md
• 37e55e4 CHANGELOG.md: reconstruct the missing entries for v1.38.1 through v1.44.1 and...
• 2f7234f AUTHORS, CONTRIBUTORS: add Nathan Herring; CHANGELOG.md: credit him by name
• See full diff in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

---

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore <dependency name> major version will close this group update PR and stop Dependabot creating any more for the specific dependency's major version (unless you unignore this specific dependency's major version or upgrade to it yourself)
• @dependabot ignore <dependency name> minor version will close this group update PR and stop Dependabot creating any more for the specific dependency's minor version (unless you unignore this specific dependency's minor version or upgrade to it yourself)
• @dependabot ignore <dependency name> will close this group update PR and stop Dependabot creating any more for the specific dependency (unless you unignore this specific dependency or upgrade to it yourself)
• @dependabot unignore <dependency name> will remove all of the ignore conditions of the specified dependency
• @dependabot unignore <dependency name> <ignore condition> will remove the ignore condition of the specified dependency and ignore conditions

[dependabot]

Looks like these dependencies are updatable in another way, so this is no longer needed.