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.
Stanford computer scientist James Zou and colleagues published Paper2Agent in Nature on September 16, an open-source framework that converts academic papers and their codebases into runnable AI agents. Fed AlphaGenome do…
BNB Chain released version 4 of BNB Agent Studio, adding NodeOps integration, stablecoin support, and fewer setup steps for deploying autonomous on-chain AI agents. The platform, first launched in July 2026 with one-prom…
Sanity shipped a September release slate headlined by Knowledge Bases in Sanity Context, which extracts facts from content, docs sites, or PDF folders and serves them to agents over an MCP endpoint, plus Workflows, a bet…
Google released AX, an open-source Apache 2.0-licensed orchestrator and declarative runtime for autonomous AI agent workloads, hosted at agentexecutor.io and on GitHub as google/ax. AX runs on Agent Substrate and treats …
A developer built a Gomoku arena at gomoku.pickuma.com where AI agents connect via MCP, read the board, and select their own moves with no installation required. In a 31-move, 21-minute match, Claude (claude-opus-5) play…
Salt Security expanded its Agentic Security Platform with native AI Detection and Response (AI-DR) capabilities that link attacks on large language models to subsequent activity across MCP servers, tools and APIs. The ne…
Firedrill released a simulation and testing framework for AI agents that lets developers define synthetic stateful tools over HTTP, MCP, CLI, and function bindings and assert on tool calls, state changes, and events. The…
Wayfair chief technology officer Fiona Tan has pushed generative AI deeper into the retailer's operations over the past year, using it to enrich a catalog of 30 million to 40 million products, clean up contradictory supp…
NetBrain Technologies has expanded its agentic automation platform to let network engineers build and deploy custom AI agents for tasks such as proactive fault identification, available to all customers from early Octobe…
ScopeTrail, a new Apache 2.0 TypeScript library, has launched to provide audit receipts for multi-hop agent delegation, minting Ed25519 signatures over JCS (RFC 8785) and storing them in an AT Protocol PDS. The library b…
Teradata Corp. is expanding its Tera AI assistant with a Context Engine, an execution system called Tera Harness and reusable Agent Skills, with the new capabilities available in the fourth quarter of 2026. In company-re…
Baselayer, officially Osiris Ratings Inc., raised $35 million in early-stage funding and launched its Agentic Identity Suite, which it calls the first interoperable trust layer and fraud prevention platform for the agent…
This+that, a workflow automation platform, is pitching LLM agents as a way for small teams to automate work that starts itself from incoming messages, citing Microsoft's finding that the average worker is interrupted onc…
Okta Inc. announced Agent Gateway, a runtime enforcement product that sits in the execution path between an AI agent and the tools it calls, plus a kill switch that revokes every active token and shuts down in-flight ses…
Wiz announced that the WIN partners endpoint in the Wiz Model Context Protocol (MCP) is now generally available, enabling partner AI agents to pull live Wiz context on demand from customer tenants. The endpoint lets part…
DigitalOcean launched Managed Agents in public preview in September 2026, combining Harness Runtime for coding harnesses in Firecracker microVMs with Action Gateway, a managed MCP endpoint offering governed access to 16,…
Bright Security detailed a dynamic application security testing (DAST) capability that detects MCP tool inventory disclosure in running LLM applications, where a user-facing assistant reveals internal runtime tool names …
A developer working on the Lakekeeper catalog built a runnable example extending zero trust to agent memory, in which two agents keep private memory in a single catalog, propose skills they cannot publish, and a human pr…
A developer has published a guide on fixing AI agent crashes caused by stateful Model Context Protocol (MCP) services, arguing that the default stdio process-spawning model breaks stateful tools like Telegram sessions, a…
A developer built a four-step "bootstrap pipeline" that indexes unfamiliar codebases by linking tests to the exact functions they execute, using a custom sys.settrace plugin to trace 1,727 tests. Measurement showed name-…