Agent Protocols news and analysis on Web Pulse: 3450 curated articles tracking the latest Agent Protocols developments, tools, and research, updated continuously from vetted sources.
A developer's guide explains that the Model Context Protocol (MCP) standardizes how AI applications discover and call external tools, resources, and prompts, but does not by itself make an agent secure. It walks through …
A developer building Pryzm Memory, a remote Model Context Protocol (MCP) server, demonstrated how connecting multiple AI assistants to a single shared memory endpoint lets context saved in one assistant be retrieved by a…
A developer has released T-Zero Context Architect V3, an open-source tool that uses AST-based static analysis to strip function implementations from a codebase while preserving signatures, docstrings, decorators and impo…
The Solon AI project released solon-ai-ui, a translation module that converts Solon AI's internal Flux<ChatEvent> stream into frontend-facing event streams for two UI protocols: the Vercel AI SDK UI Message Stream Protoc…
A developer has published a production playbook for hardening Model Context Protocol (MCP) servers against prompt injection and tool abuse in .NET, arguing that server-side policy enforcement beats client-side sanitizati…
JetBrains released Air in IDEs in Early Access on October 1, 2026, shipping its agentic orchestration layer as a free plugin inside IntelliJ, Rider, PhpStorm, WebStorm, and PyCharm rather than as a separate product. Air …
A developer outlines how agent skills, packaged as text-based SKILL.md contracts in ecosystems like OpenClaw and its ClawHub registry, are emerging as an installable integration layer that lets AI assistants read data an…
Anthropic announced its Model Hardware Standard (MHS) at the end of August as an open interface for connecting AI models and agents to hardware, and received 30x more organizations requesting to participate in the first …
TypeSafe's jev-browser-wingman MCP proxy completed all 34 browser-task runs in a 68-run comparison against Playwright alone, which completed 32 of 34 (94%), with both Playwright misses on the file-upload task. The proxy'…
OX Security disclosed in April 2026 that a systemic flaw in the STDIO transport of Anthropic's Model Context Protocol lets a malicious MCP server execute arbitrary operating system commands with no sanitization, exposing…
Claude Code has read AGENTS.md files since version 2.1.277, released 2026-09-18, and since version 2.1.281, released 2026-09-23, does so in every session including Amazon Bedrock, Google Vertex AI, Microsoft Foundry, LLM…
Writing rules once in a repo-root AGENTS.md file lets Codex and Cursor read them directly, while Claude Code reads AGENTS.md only when no CLAUDE.md exists or via an @AGENTS.md import line, according to a blog post publis…
A six-point self-audit published by an unnamed technical blog gives developers pass/fail curl checks for the files AI agents from Claude, ChatGPT and Gemini read to discover and call services, covering llms.txt, the A2A …
Sierra announced the Personal Agent Protocol (PAP) on October 6, 2026, an open standard co-developed with Meta for connecting personal AI agents to businesses, with Genesys, Instinct, Rocket, Shopify, Stripe and Walmart …
OpenAI's Codex loads skills from a SKILL.md file in .agents/skills folders, capping the initial skill list at 2 percent of the model's context window or 8,000 characters when the window is unknown, according to OpenAI's …
Pi 1.0 added MCP support through a feature called Codemode, which runs JavaScript in a QuickJS sandbox inside a WASM runtime on the harness side with no network, no file system, no timers and limited RAM, allowing an LLM…
Remote MCP server deployments most commonly fail with a 401 because of OAuth discovery problems rather than token problems, according to a debugging guide covering the four-step discovery chain defined by RFC 9728 protec…
Sierra announced the Personal Agent Protocol (PAP) on October 6, 2026, an open standard defining how a person's AI agent connects to a business, with Meta co-developing and Genesys, Instinct, Rocket, Shopify, Stripe and …
A team operating a WooCommerce store traced a cart-abuse campaign to 8,096 unique IP addresses rotating through browser-like Chrome/150 user agents, with most addresses appearing only once or twice. Rather than block IPs…
Independent researcher Syed Anas Mohiuddin developed proof-of-concept attacks that exploit trust gaps in the Model Context Protocol (MCP) to turn one compromised AI agent into a vector for spreading malicious instruction…