MCP
mcp-393fa6aa·13 events·first seen 15d agoAliases: MCP
Co-occurring entities
More like this (12)
Recent events (13)
Hugging Face integrates MCP tools with Reachy Mini robot
Hugging Face published a blog post describing how to add Model Context Protocol (MCP) tools to the Reachy Mini robot platform. The integration connects MCP-based tool-calling infrastructure to physical robotics hardware. This is a concrete deployment example of MCP expanding beyond software agents into embodied AI systems.
HexStrike AI: MCP server exposing 150+ cybersecurity tools to AI agents
HexStrike AI is an open-source MCP server that enables AI agents (Claude, GPT, Copilot, and others) to autonomously invoke over 150 offensive security tools for penetration testing, vulnerability discovery, and bug bounty automation. The project bridges LLMs with real-world offensive security capabilities via the Model Context Protocol. With 9,221 GitHub stars, it represents a notable community signal around agentic security tooling and the expanding attack surface of AI-driven automation.
AgentBeats: Standardized Agent Evaluation via A2A and MCP Protocols
A new arXiv preprint proposes Agentified Agent Assessment (AAA), a framework where evaluation is performed by judge agents interacting through standardized protocols—A2A for task management and MCP for tool access—rather than bespoke benchmark harnesses. The authors introduce AgentBeats as a concrete implementation, validated through a five-month open competition with 298 judge agents and 467 subject agents across 12 categories, plus a coding-agent case study. The work addresses fragmentation in agent evaluation by decoupling assessment logic from agent implementation, enabling reproducible and interoperable benchmarking.
IBM releases mcp-context-forge: AI gateway and proxy for MCP, A2A, and REST/gRPC APIs
IBM has open-sourced mcp-context-forge, a Python-based AI gateway, registry, and proxy that sits in front of MCP, A2A, or REST/gRPC APIs and exposes a unified endpoint with centralized discovery, guardrails, and management. The tool is designed to optimize agent and tool calling workflows and supports plugins. With ~3,800 GitHub stars, it represents a notable infrastructure contribution to the MCP/A2A ecosystem from a major enterprise vendor.
DesktopCommanderMCP: MCP server giving Claude terminal control and file system capabilities
DesktopCommanderMCP is an open-source TypeScript MCP server that extends Claude with terminal control, file system search, and diff-based file editing capabilities. The project has accumulated 6,137 GitHub stars, indicating meaningful community adoption. It represents a practical implementation of the MCP protocol for agentic desktop control use cases.
LibreChat: open-source multi-provider chat interface with agents and MCP support
LibreChat is an open-source, self-hostable ChatGPT-style interface supporting a wide range of model providers including OpenAI, Anthropic, DeepSeek, Mistral, Google Vertex AI, and others. It includes agent capabilities, MCP (Model Context Protocol) integration, code interpreter, DALL-E-3, and multi-user authentication. The project has accumulated over 38,600 GitHub stars with active daily growth, indicating significant community adoption.
SkillWeaver: Compositional Skill Routing for LLM Agents via Decompose-Retrieve-Compose
Researchers introduce SkillWeaver, a framework for compositional skill routing in LLM agents that decomposes complex queries into atomic sub-tasks, retrieves matching skills from a large library, and composes an executable DAG plan. The paper formalizes the Compositional Skill Routing problem and introduces CompSkillBench, a benchmark of 300 compositional queries over 2,209 real MCP server skills across 24 categories. A key finding is that task decomposition quality is the primary bottleneck, with standard LLM decomposition reaching only 34.2% category recall; the proposed Iterative Skill-Aware Decomposition (SAD) method improves decomposition accuracy from 51.0% to 67.7% in a single iteration. The framework also reduces context window consumption by over 99% compared to naive skill-stuffing approaches.
MiniMax M2.7 proprietary reasoning model competes with Gemini and Claude Opus; roundup covers Cursor Composer 2, MAI-Image-2, Claude Code Channels, and Anthropic defense dispute
MiniMax released M2.7, a proprietary reasoning model that achieved 66.6% on MLE Bench Lite (tying Gemini 3.1) and 56.22% on SWE-Pro, priced at $0.30/$1.20 per million tokens, with the shift to proprietary marking a potential strategic pivot among Chinese AI labs away from open weights. Cursor released Composer 2, an agentic coding model built on a fine-tuned Kimi 2.5 (via Moonshot partnership), priced 86% cheaper than its predecessor and scoring 73.7 on SWE-bench Multilingual. Anthropic released Claude Code Channels, routing Telegram and Discord messages into local Claude Code sessions via MCP plugins, and separately filed a court response denying it has any backdoor or kill switch into military deployments of Claude. Microsoft announced MAI-Image-2, a text-to-image model ranking third on Arena.ai among research labs.
Yuxi: Multi-tenant agent harness integrating LightRAG, knowledge graphs, and MCP
Yuxi is an open-source multi-tenant agent harness platform that combines a LightRAG knowledge base with knowledge graph management. Built on LangChain, Vue, and FastAPI, it supports DeepAgents, MinerU PDF parsing, Neo4j, and the Model Context Protocol (MCP). The project has accumulated 5,451 GitHub stars with modest daily traction (+47).
CHAP: Collaborative Human-Agent Protocol for structured human-AI accountability in multi-agent deployments
Researchers from BrightbeamAI introduce CHAP (Collaborative Human-Agent Protocol), a protocol specification for formalizing human-agent collaboration in production multi-agent systems. CHAP defines shared workspaces, structured override events with diffs and rationales, non-repudiable signed approvals, and an append-only evidence log, filling a gap left by MCP (tool access) and A2A (agent-to-agent interoperability). The protocol ships with a reference implementation, conformance suite, and worked examples. It targets high-stakes deployments in domains like clinical decisions, contracts, and code where human judgment must be auditable and replayable.
Claude Code and New Admin Controls Available for Anthropic Team and Enterprise Plans
Anthropic is expanding Claude Code access to Team and Enterprise customers via premium seat upgrades, bundling the coding agent with Claude under a single subscription. New admin controls include granular spend caps, self-serve seat management, managed policy settings for Claude Code (including MCP server configurations), and usage analytics. A new Compliance API provides programmatic real-time access to usage data and customer content for auditing, monitoring, and data retention management. Early enterprise customers Behavox and Altana report significant productivity gains.
Anthropic launches Services Track tiers and Partner Hub for Claude Partner Network
Anthropic announced two additions to its Claude Partner Network: a Services Track with three tiers (Select, Preferred, Global Premier) based on certified practitioners, production deployments, and public customer references, and a Claude Partner Hub portal where partners track their standing and customers find qualified implementation firms. The program, backed by a $100M investment, has attracted over 40,000 applicants and 10,000 certified consultants since its March 2026 launch. Major professional services firms including Accenture, Cognizant, Deloitte, KPMG, Infosys, and PwC are building Claude practices at scale, with some rolling out access to hundreds of thousands of employees. A new MCP connector allows partners to query their partnership status directly through Claude.
basic-memory: persistent memory layer for AI conversations via MCP
basic-memory is an open-source Python project providing persistent memory for AI conversations, allowing users to avoid re-explaining context across sessions. The project has accumulated 3,212 GitHub stars with modest daily growth. It appears to implement a local knowledge-graph or note-based memory store exposed via MCP or similar tooling.