Agent Protocols news and analysis on Web Pulse: 3470 curated articles tracking the latest Agent Protocols developments, tools, and research, updated continuously from vetted sources.
Snyk announced the Evo MCP server on September 22, 2026, a remote server that lets developers query AI asset inventory, risk data, policies, and open violations from inside agent harnesses such as Cursor and Claude Code.…
A developer working on MCPFusion, the open-source TypeScript framework behind Vinkius connectors, built an MCP-based integration that gives AI agents governed access to Bright Data's proxy and web-unlocking infrastructur…
A developer argues that the Model Context Protocol (MCP), while easy to run locally, hits an "Enterprise MCP Wall" at scale, where hardcoded credentials, unregistered server sprawl, and overly broad tool permissions crea…
Aster Market, an early marketplace focused on bounded digital services, is arguing that AI-agent marketplaces should prioritize verifiable delivery over listing volume. The project proposes that listings be bounded, mach…
A developer released AIFeed, a protocol that lets publishers sign a small JSON manifest with Ed25519 and anchor the key in DNS so AI crawlers can cryptographically verify content permissions rather than relying on robots…
BlindPay published version 1.8.0 of its official MCP server to the official MCP registry on 2026-09-22, distributed as @blindpay/mcp on npm and as a hosted remote over streamable-http for stablecoin payouts, payins, quot…
The HaloPSA MCP server, version 1.7.15, was published to the official MCP registry on 2026-09-22 as the Docker image ghcr.io/wyre-ai/halopsa-mcp:v1.7.15, according to the mcpindex server index. The server, which connects…
The ConnectWise Automate MCP server, published by WYRE-AI as version 2.0.2 to the official MCP registry on 2026-09-22, passed a semantic description screen with an advisory posture, according to the mcpindex index. The s…
Rails AI Context version 5.27.0, an MCP server providing 45 MCP tools that give AI agents live Rails schema, routes, models, and conventions, was published to the official MCP registry on 2026-09-22, according to the mcp…
KitchenSink4PPT, a PowerPoint (.pptx) MCP server published as kitchensink4ppt on PyPI, reached version 1.3.1 in the official MCP registry on 2026-09-22, according to the mcpindex index. The index's semantic screen of the…
Zero Slop v2.12.2, an MCP server that scores and rewrites AI-assisted prose while preserving names, numbers, quotes, links, and facts, was published to the official MCP registry on 2026-09-22 and is distributed as a host…
The io.github.mrfentmen/discord-mcp MCP server, a keyless tool for Discord invites and server widgets that requires no bot token, was published to the official MCP registry on 2026-09-22 as version 1.0.0 and distributed …
Notion launched the Notion Skills API, a set of API endpoints for loading skills from Notion in spec-compliant file formats, and open-sourced a starter kit for syncing team skills from Notion into GitHub. The API is desi…
A blog post on maharship.com arguing that the Model Context Protocol (MCP) was designed for 2024-era models and now imposes context bloat and a token tax sparked a roughly 165-point Hacker News thread with over 100 comme…
AI coding agents can produce changes that pass TypeScript, tests, and builds yet still violate a project's intended architecture, according to a first-person account of working with coding agents on larger changes. The a…
Agent-shell released version 0.78, adding a persistent writable prompt, mid-turn prompt queueing and steering, and two new agents, Antigravity and Qoder, to the Emacs mode for interacting with AI agents over the Agent Cl…
A solo developer built Reset Beer, a dashboard that aggregates official status pages from more than 15 AI products and alerts users when quota reset signals appear. The project also ships an MCP server, distributed as th…
A developer released Context Tetris, a game in which players clear full rows to collect their contents and bank 4 instructions, 4 documents and 4 memories to answer a task. Noise fills space but earns no context, while c…
NeuralTrust is a runtime security and governance layer for enterprise AI agents that secures connections to models, tools, and MCP servers through a gateway, per the vendor's site at neuraltrust.ai. The platform, aimed a…
Fastly integrated three agent-specific capabilities — an AI Firewall, AI Runtime Control, and expanded API Security — into its global edge cloud, which carries 622 Tbps of capacity and processes over 5 trillion requests …