llm-kb
← Back to releases
Tool Release

Model Context Protocol releases v2026.7.4 with updated server packages

The Model Context Protocol project published version v2026.7.4 on July 4, 2026, updating four server packages: everything, filesystem, sequential-thinking, and memory. This routine release likely includes bug fixes and improvements for MCP server implementations.

0 engagement·1 source·Sat, Jul 4, 2026, 11:05 PM
The release, tagged as v2026.7.4 on GitHub, updates four packages under the @modelcontextprotocol scope: server-everything, server-filesystem, server-sequential-thinking, and server-memory. No changelog or specific fixes are mentioned in the post. The release is part of the ongoing maintenance of the Model Context Protocol, which provides standardized server implementations for AI agent interactions.

Entities

server-everything(tool)server-filesystem(tool)server-sequential-thinking(tool)server-memory(tool)@modelcontextprotocol/server-filesystem(tool)Model Context Protocol(tool)@modelcontextprotocol/server-everything(tool)@modelcontextprotocol/server-sequential-thinking(tool)@modelcontextprotocol/server-memory(tool)

Related

Tool ReleaseFri, Jul 10, 2026, 01:45 AM

MCP servers and Claude Code CLI updated with new features and fixes

On July 10, 2026, the Model Context Protocol (MCP) servers and Claude Code CLI received updates. The MCP servers release v2026.7.10 updated packages for filesystem, time, fetch, and git. Claude Code CLI v2.1.206 added directory path suggestions for /cd, a /doctor check for trimming CLAUDE.md files, and improved git push handling.

14 engagement·2 sources·github
CommunitySun, Jul 12, 2026, 01:21 PM

Only 1 of 4,356 reachable MCP servers is ready for the 2026-07-28 spec

A scan of 4,356 reachable MCP servers found that only one is compliant with the upcoming 2026-07-28 specification. This highlights a severe lack of readiness across the MCP ecosystem just two weeks before the spec deadline.

14 engagement·1 source·hackernews
Tool ReleaseThu, Jul 9, 2026, 05:56 PM

LangChain Fireworks partner package v1.4.4 released with cached prompt token reporting

LangChain released version 1.4.4 of the langchain-fireworks partner package on July 9, 2026. The update adds reporting of cached prompt token usage and includes dependency bumps and documentation clarifications.

0 engagement·1 source·github
Tool ReleaseSun, Jul 5, 2026, 08:53 PM

LangChain-OpenRouter releases v0.2.6 with custom HTTP header support

LangChain-OpenRouter v0.2.6 is released on 2026-07-05, adding support for `default_headers` for custom HTTP header injection and refreshing model profile data. The update is a minor version bump from 0.2.5.

0 engagement·1 source·github
Tool ReleaseSat, Jul 11, 2026, 08:06 PM

vLLM v0.25.0 released with Model Runner V2 as default and PagedAttention removed

vLLM v0.25.0 is now available, featuring 558 commits from 232 contributors. Model Runner V2 becomes the default execution path for all dense models, with new support for EVS, realtime embeddings, prefix caching for Mamba hybrid models, multimodal-prefix bidirectional attention, and dynamic speculative decoding. The legacy PagedAttention implementation has been removed.

0 engagement·1 source·github