Agent Protocols news and analysis on Web Pulse: 3381 curated articles tracking the latest Agent Protocols developments, tools, and research, updated continuously from vetted sources.
A developer released `dirigista`, an MCP server that exposes three classification tools — `verify`, `categorize` and `score` — to MCP clients including Claude Code, Claude Desktop and Cursor, returning typed judgments wi…
A new Agentic AI red team checklist released by CyberUpdates365 provides 222 structured security tests across 20 categories, extending assessments beyond prompt injection into cloud, tools, memory, MCP and agent-to-agent…
Google Cloud announced at its Gemini at Work 2026 event on October 8 an enterprise AI agent that holds its own Workspace identity, including a dedicated email address, calendar, Drive storage, and a directory listing und…
Modelship, an open source Ray-backed model cluster, was released on GitHub, scoring 17/17 on the Open Responses conformance suite according to its production readiness documentation. The project spins up a Ray cluster, d…
A developer built Browser Cookie Bridge, a macOS app that transfers cookies from Chromium browsers like Brave into AI coding agents' in-app browsers so agents can start from existing logged-in sessions. After hitting RAM…
A developer released mcp-era-probe v0.8, a read-only tool that probes how Model Context Protocol servers respond when a client calls a nonexistent tool, and found servers answer in three structurally different ways: a -3…
A developer published a comparison of open protocols for multi-agent systems, contrasting Anthropic's MCP for agent-to-tool connections, Google's A2A for agent-to-agent delegation, and alternatives including ACP, ANP, an…
A developer released a11yscope-mcp, an MCP server that wraps their WCAG accessibility checker so coding agents like Claude Code and Cursor can scan and fix generated HTML. The server exposes three tools — scan_page, scan…
A developer localized the README of Helix Foundry, an Apache-2.0 TypeScript project (1029 stars) that connects a company's existing databases, Stripe, WorkOS, PostHog, files and generic APIs into a cross-source ontology …
Developer vscarpenter released GSD Task Manager version 13.9.2, an offline-first Eisenhower matrix task app that ships with a separately installed MCP server exposing 19 task, analytics, and diagnostic tools to AI agents…
A developer built a Model Context Protocol server exposing two decades of Indian stock index data and found that an AI agent's correct answer — 23 March 2010 as the worst 10-year buy day for all three Nifty indices — loo…
A developer's 14 headless test runs of Claude Code 2.1.289 found that a hook matcher of `mcp__lab` fired 0 times across 42 calls to that MCP server's tools, while `mcp__lab__.*` fired 42 of 42, even though the same bare …
Jshookmcp released an MCP server that instruments a real signed-in Chrome browser over CDP for front-end reverse engineering, exposing 735 tools across 36 self-discovered domains. The server's search profile loads about …
Developer Herval released littleguys, an open-source macOS menu bar app that monitors and controls Claude Code, Codex, Gemini, OpenClaw and Hermes agents across multiple machines. The app speaks the Agent Client Protocol…
IHMT-MEMORY, a new long-term memory system for AI coding agents, stores decisions, preferences and corrections as a tree of plain text files on the user's disk and retrieves typically 200–900 tokens per query whether the…
A developer's analysis of computer-use agents in production argues that MCP (Model Context Protocol) occupies a narrow slice of the stack between the model's reasoning loop and tool execution, and that agent-to-agent com…
Atlassian on October 6, 2026 described its Agentic Multiplayer Protocol (AMP) as the basis for people and agents to work together inside its platform, built on an identity you assign, authority you scope, shared context …
Anthropic added dynamic workflows to Claude Managed Agents in beta on 9 Oct 2026, letting an agent write a program that runs many agents in phases and combines their results, with a limit of 1,000 agents started per run.…
REA is a Model Context Protocol server that connects AI coding agents to established reverse engineering tools — Ghidra (released by the NSA in 2019), Hopper (2011), and IDA (1990s) — so an agent can decompile an app and…
A developer built handup, an inbox that routes questions, approval requests and completion notices from multiple coding agents (Claude Code, Codex and others) to a single desktop or phone interface, letting the user appr…