Agent Protocols news and analysis on Web Pulse: 3474 curated articles tracking the latest Agent Protocols developments, tools, and research, updated continuously from vetted sources.
Bruin's open-source CLI now provides three commands — `bruin ai enhance`, `bruin lineage`, and `bruin validate` — for building a git-versioned context layer that gives AI agents asset descriptions, dependency lineage, an…
A scan of 135 live x402-payable services from Coinbase's public Bazaar catalog found that 33% have no SPF record, 63% are not actively enforcing DMARC, and 27% show no agent-discoverability signal such as llms.txt, MCP, …
Croft published a blog post on September 5, 2026 arguing that a single AI assistant with tools and loadable skills outperforms a swarm of specialist subagents for building internal apps, citing Claude's platform team's r…
Agent Skills are reusable packages that define how an AI agent performs a specialised class of work, distinct from agents, prompts, tools, MCP, and workflows, according to an analysis of the six layers. The framework ass…
Synsugar GmbH released Klaaro, a schema-first document-extraction API that classifies an uploaded file, generates its own JSON schema for that document class, then fills it, with Python and TypeScript SDKs both published…
Agent-native CRM is defined by five verifiable architectural properties — full agent parity with the human UI, an agent-specific safety model, identical self-hosted and cloud agent capability, machine-readable product su…
Common Crawl's July 2026 crawl (CC-MAIN-2026-30) recorded 598,298 responses to /llms.txt and /llms-full.txt seed requests, of which 14,191 were empty, leaving 584,107 files analyzed for spec compliance, generating softwa…
WordPress.com added ChatGPT as an MCP-compatible AI agent that can draft posts, categorize content, moderate comments, fix missing alt text, and assist with SEO, with MCP access included in paid plans and a 30-day free t…
Relaticle released an MCP server at mcp.relaticle.com that lets Claude search, read, create, update, and soft-delete companies, people, opportunities, tasks, and notes in a single authorized workspace, with OAuth 2.1 and…
A practice exam has been published for the Claude Certified Architect – Foundations (CCA-F/CCAR-F) certification, covering agentic architecture and orchestration with the Claude Agent SDK. The questions test core agent-l…
Nearform built a multi-agent research system that answers natural-language questions across an OpenSearch cluster indexing over 330,000 documents across roughly 30 indexes, after testing the OpenSearch MCP server and que…
A condensed eight-week AI engineering curriculum has been published, cutting an original 13-week programme down to the sessions deemed most valuable. Each week pairs roughly 60-90 minutes of reading and video with a hand…
WebMCP, a proposed browser standard incubated at the W3C by Google's Chrome team and Microsoft's Edge team, lets websites expose structured tools that AI agents call directly instead of screen-scraping pages, and it has …
A developer has reconstructed the deployed architecture of Grok Bot, a stateful agent service that runs inside a Linux container, from recovered JavaScript bundles and runtime scripts. The reconstruction, covering 1,804 …