diff --git a/.changeset/576-streamable-http.md b/.changeset/576-streamable-http.md deleted file mode 100644 index 9e6d847bb7..0000000000 --- a/.changeset/576-streamable-http.md +++ /dev/null @@ -1,7 +0,0 @@ ---- -"@modelcontextprotocol/server-filesystem": minor -"@modelcontextprotocol/server-memory": minor -"@modelcontextprotocol/server-sequential-thinking": minor ---- - -Serve Streamable HTTP as well as stdio (#576). `--transport streamable-http` serves the server on one `/mcp` endpoint, for 2026-07-28 and 2025-era clients alike, at `HOST`:`PORT` (`127.0.0.1:3001` by default); stdio stays the default. Over HTTP the filesystem server requires its allowed directories on the command line, since it cannot ask a client for its Roots, and the sequential-thinking history belongs to the server process rather than to one connection. diff --git a/.changeset/ghsa-38j4-j742-j7r3.md b/.changeset/ghsa-38j4-j742-j7r3.md deleted file mode 100644 index fe3a684250..0000000000 --- a/.changeset/ghsa-38j4-j742-j7r3.md +++ /dev/null @@ -1,5 +0,0 @@ ---- -"@modelcontextprotocol/server-everything": patch ---- - -Fix GHSA-38j4-j742-j7r3. diff --git a/.changeset/ghsa-rq2p-q7c4-6g27.md b/.changeset/ghsa-rq2p-q7c4-6g27.md deleted file mode 100644 index acb9765c55..0000000000 --- a/.changeset/ghsa-rq2p-q7c4-6g27.md +++ /dev/null @@ -1,5 +0,0 @@ ---- -"@modelcontextprotocol/server-filesystem": patch ---- - -Fix GHSA-rq2p-q7c4-6g27. diff --git a/.changeset/memory-lock-exit-cleanup.md b/.changeset/memory-lock-exit-cleanup.md deleted file mode 100644 index 54e857fdb8..0000000000 --- a/.changeset/memory-lock-exit-cleanup.md +++ /dev/null @@ -1,5 +0,0 @@ ---- -"@modelcontextprotocol/server-memory": patch ---- - -Remove the memory file lock the server holds when it is stopped by SIGINT or SIGTERM, or exits, so the next writer does not wait for it to go stale. diff --git a/.changeset/memory-stale-lock-successor.md b/.changeset/memory-stale-lock-successor.md deleted file mode 100644 index e306469fd5..0000000000 --- a/.changeset/memory-stale-lock-successor.md +++ /dev/null @@ -1,5 +0,0 @@ ---- -"@modelcontextprotocol/server-memory": patch ---- - -Stop a server breaking a stale memory file lock from deleting a live lock that replaced it in the meantime, which let two writers change the memory file at once and lose a write. The stale lock is now moved aside, checked to be the file that was judged stale, and only then deleted. diff --git a/.changeset/memory-write-limits.md b/.changeset/memory-write-limits.md deleted file mode 100644 index 1d09bb04ed..0000000000 --- a/.changeset/memory-write-limits.md +++ /dev/null @@ -1,5 +0,0 @@ ---- -"@modelcontextprotocol/server-memory": minor ---- - -Add opt-in write limits: `MEMORY_MAX_ENTITIES`, `MEMORY_MAX_RELATIONS` and `MEMORY_MAX_FILE_BYTES` cap how large the write tools may grow the graph, and the README now documents that every write rewrites the whole memory file. diff --git a/.changeset/sequentialthinking-header-row-cap.md b/.changeset/sequentialthinking-header-row-cap.md deleted file mode 100644 index a753cedf67..0000000000 --- a/.changeset/sequentialthinking-header-row-cap.md +++ /dev/null @@ -1,5 +0,0 @@ ---- -"@modelcontextprotocol/server-sequential-thinking": patch ---- - -Count the thought box's header rows toward its 1000-row cap, so a long `branchId` can no longer draw past it on stderr: line breaks in a `branchId` are shown as `\r` and `\n`, and a header too long for the cap is cut with a row saying how many characters it left out (#5526). diff --git a/package-lock.json b/package-lock.json index d31f0ae844..b0d0cd9f9f 100644 --- a/package-lock.json +++ b/package-lock.json @@ -5253,7 +5253,7 @@ }, "src/everything": { "name": "@modelcontextprotocol/server-everything", - "version": "2.3.0", + "version": "2.3.1", "license": "SEE LICENSE IN LICENSE", "dependencies": { "@modelcontextprotocol/node": "^2.1.1", @@ -5282,7 +5282,7 @@ }, "src/filesystem": { "name": "@modelcontextprotocol/server-filesystem", - "version": "2.3.0", + "version": "2.4.0", "license": "SEE LICENSE IN LICENSE", "dependencies": { "@modelcontextprotocol/node": "^2.1.1", @@ -5310,7 +5310,7 @@ }, "src/memory": { "name": "@modelcontextprotocol/server-memory", - "version": "2.1.1", + "version": "2.2.0", "license": "SEE LICENSE IN LICENSE", "dependencies": { "@modelcontextprotocol/node": "^2.1.1", @@ -5334,7 +5334,7 @@ }, "src/sequentialthinking": { "name": "@modelcontextprotocol/server-sequential-thinking", - "version": "2.1.1", + "version": "2.2.0", "license": "SEE LICENSE IN LICENSE", "dependencies": { "@modelcontextprotocol/node": "^2.1.1", diff --git a/src/everything/CHANGELOG.md b/src/everything/CHANGELOG.md index c2c9b47988..1ce679f52a 100644 --- a/src/everything/CHANGELOG.md +++ b/src/everything/CHANGELOG.md @@ -1,5 +1,11 @@ # @modelcontextprotocol/server-everything +## 2.3.1 + +### Patch Changes + +- [#5583](https://github.com/modelcontextprotocol/servers/pull/5583) [`66f641e`](https://github.com/modelcontextprotocol/servers/commit/66f641e9ae2fc7726a9420b15fbf68e0613a254b) [@cliffhall](https://github.com/cliffhall) - Fix GHSA-38j4-j742-j7r3. + ## 2.3.0 ### Minor Changes diff --git a/src/everything/package.json b/src/everything/package.json index 2beb5f5648..9d2155fc3f 100644 --- a/src/everything/package.json +++ b/src/everything/package.json @@ -1,6 +1,6 @@ { "name": "@modelcontextprotocol/server-everything", - "version": "2.3.0", + "version": "2.3.1", "description": "MCP server that exercises all the features of the MCP protocol", "license": "SEE LICENSE IN LICENSE", "mcpName": "io.github.modelcontextprotocol/server-everything", diff --git a/src/filesystem/CHANGELOG.md b/src/filesystem/CHANGELOG.md index be32e0b302..c48aaf6ccc 100644 --- a/src/filesystem/CHANGELOG.md +++ b/src/filesystem/CHANGELOG.md @@ -1,5 +1,15 @@ # @modelcontextprotocol/server-filesystem +## 2.4.0 + +### Minor Changes + +- [#5536](https://github.com/modelcontextprotocol/servers/pull/5536) [`311c0fd`](https://github.com/modelcontextprotocol/servers/commit/311c0fdc953f38da08153fd2e8d134d03c8e168b) [@cliffhall](https://github.com/cliffhall) - Serve Streamable HTTP as well as stdio ([#576](https://github.com/modelcontextprotocol/servers/issues/576)). `--transport streamable-http` serves the server on one `/mcp` endpoint, for 2026-07-28 and 2025-era clients alike, at `HOST`:`PORT` (`127.0.0.1:3001` by default); stdio stays the default. Over HTTP the filesystem server requires its allowed directories on the command line, since it cannot ask a client for its Roots, and the sequential-thinking history belongs to the server process rather than to one connection. + +### Patch Changes + +- [#5584](https://github.com/modelcontextprotocol/servers/pull/5584) [`89b1152`](https://github.com/modelcontextprotocol/servers/commit/89b1152805e832371219d27fb9ef885e5a325dfb) [@cliffhall](https://github.com/cliffhall) - Fix GHSA-rq2p-q7c4-6g27. + ## 2.3.0 ### Minor Changes diff --git a/src/filesystem/package.json b/src/filesystem/package.json index 56a4d744d1..63c6d015ac 100644 --- a/src/filesystem/package.json +++ b/src/filesystem/package.json @@ -1,6 +1,6 @@ { "name": "@modelcontextprotocol/server-filesystem", - "version": "2.3.0", + "version": "2.4.0", "description": "MCP server for filesystem access", "license": "SEE LICENSE IN LICENSE", "mcpName": "io.github.modelcontextprotocol/server-filesystem", diff --git a/src/memory/CHANGELOG.md b/src/memory/CHANGELOG.md index 70aa2bdac7..899f96b267 100644 --- a/src/memory/CHANGELOG.md +++ b/src/memory/CHANGELOG.md @@ -1,5 +1,19 @@ # @modelcontextprotocol/server-memory +## 2.2.0 + +### Minor Changes + +- [#5536](https://github.com/modelcontextprotocol/servers/pull/5536) [`311c0fd`](https://github.com/modelcontextprotocol/servers/commit/311c0fdc953f38da08153fd2e8d134d03c8e168b) [@cliffhall](https://github.com/cliffhall) - Serve Streamable HTTP as well as stdio ([#576](https://github.com/modelcontextprotocol/servers/issues/576)). `--transport streamable-http` serves the server on one `/mcp` endpoint, for 2026-07-28 and 2025-era clients alike, at `HOST`:`PORT` (`127.0.0.1:3001` by default); stdio stays the default. Over HTTP the filesystem server requires its allowed directories on the command line, since it cannot ask a client for its Roots, and the sequential-thinking history belongs to the server process rather than to one connection. + +- [#5534](https://github.com/modelcontextprotocol/servers/pull/5534) [`1c00bfa`](https://github.com/modelcontextprotocol/servers/commit/1c00bfa0f3a38faee6be46631f95928726e37ab1) [@cliffhall](https://github.com/cliffhall) - Add opt-in write limits: `MEMORY_MAX_ENTITIES`, `MEMORY_MAX_RELATIONS` and `MEMORY_MAX_FILE_BYTES` cap how large the write tools may grow the graph, and the README now documents that every write rewrites the whole memory file. + +### Patch Changes + +- [#5530](https://github.com/modelcontextprotocol/servers/pull/5530) [`37bff57`](https://github.com/modelcontextprotocol/servers/commit/37bff572435208d09b5b7403edf638c88871cc14) [@cliffhall](https://github.com/cliffhall) - Remove the memory file lock the server holds when it is stopped by SIGINT or SIGTERM, or exits, so the next writer does not wait for it to go stale. + +- [#5563](https://github.com/modelcontextprotocol/servers/pull/5563) [`d57917b`](https://github.com/modelcontextprotocol/servers/commit/d57917bc0661f7342a50272be572e04f8962a6ca) [@cliffhall](https://github.com/cliffhall) - Stop a server breaking a stale memory file lock from deleting a live lock that replaced it in the meantime, which let two writers change the memory file at once and lose a write. The stale lock is now moved aside, checked to be the file that was judged stale, and only then deleted. + ## 2.1.1 ### Patch Changes diff --git a/src/memory/package.json b/src/memory/package.json index 3a6c17701a..5951582337 100644 --- a/src/memory/package.json +++ b/src/memory/package.json @@ -1,6 +1,6 @@ { "name": "@modelcontextprotocol/server-memory", - "version": "2.1.1", + "version": "2.2.0", "description": "MCP server for enabling memory for Claude through a knowledge graph", "license": "SEE LICENSE IN LICENSE", "mcpName": "io.github.modelcontextprotocol/server-memory", diff --git a/src/sequentialthinking/CHANGELOG.md b/src/sequentialthinking/CHANGELOG.md index 9700d8625e..2ca7cabe93 100644 --- a/src/sequentialthinking/CHANGELOG.md +++ b/src/sequentialthinking/CHANGELOG.md @@ -1,5 +1,15 @@ # @modelcontextprotocol/server-sequential-thinking +## 2.2.0 + +### Minor Changes + +- [#5536](https://github.com/modelcontextprotocol/servers/pull/5536) [`311c0fd`](https://github.com/modelcontextprotocol/servers/commit/311c0fdc953f38da08153fd2e8d134d03c8e168b) [@cliffhall](https://github.com/cliffhall) - Serve Streamable HTTP as well as stdio ([#576](https://github.com/modelcontextprotocol/servers/issues/576)). `--transport streamable-http` serves the server on one `/mcp` endpoint, for 2026-07-28 and 2025-era clients alike, at `HOST`:`PORT` (`127.0.0.1:3001` by default); stdio stays the default. Over HTTP the filesystem server requires its allowed directories on the command line, since it cannot ask a client for its Roots, and the sequential-thinking history belongs to the server process rather than to one connection. + +### Patch Changes + +- [#5535](https://github.com/modelcontextprotocol/servers/pull/5535) [`a5c60fa`](https://github.com/modelcontextprotocol/servers/commit/a5c60fa8cb442de81b99ccc4bcac8cd79e8c044f) [@cliffhall](https://github.com/cliffhall) - Count the thought box's header rows toward its 1000-row cap, so a long `branchId` can no longer draw past it on stderr: line breaks in a `branchId` are shown as `\r` and `\n`, and a header too long for the cap is cut with a row saying how many characters it left out ([#5526](https://github.com/modelcontextprotocol/servers/issues/5526)). + ## 2.1.1 ### Patch Changes diff --git a/src/sequentialthinking/package.json b/src/sequentialthinking/package.json index 71e91c2133..8a650518e5 100644 --- a/src/sequentialthinking/package.json +++ b/src/sequentialthinking/package.json @@ -1,6 +1,6 @@ { "name": "@modelcontextprotocol/server-sequential-thinking", - "version": "2.1.1", + "version": "2.2.0", "description": "MCP server for sequential thinking and problem solving", "license": "SEE LICENSE IN LICENSE", "mcpName": "io.github.modelcontextprotocol/server-sequential-thinking",