AI Agents news and analysis on Web Pulse: 13524 curated articles tracking the latest AI Agents developments, tools, and research, updated continuously from vetted sources.
Researchers introduced CAGE-1, an evaluation framework for enterprise agentic AI that assesses control, assurance, and governance before deployment. The framework tests authority, policy enforcement, retrieval quality, m…
Researchers introduced SkillOpt-Lite, a minimal skill optimization pipeline for autonomous agents that accelerates convergence and improves performance, achieving +8.8 points on LiveMath with GPT-5.5 and +25.4 points wit…
Theo Browne published a guide for running GPT-5.5 as a subagent within Claude Code workflows, ranking models by cost, intelligence, and taste. The approach uses GPT-5.5 for bulk mechanical work via the Codex CLI, while r…
A developer explores balancing Claude 4.8 and GLM 5.2 in structuring prompts for mid-level coding agents, aiming to optimize performance in automated code generation tasks.
Anthropic released Claude Fable 5, a new AI model that improves agentic coding by helping developers discover their unknowns. The model's effectiveness depends on users' ability to clarify their intent, making the skill …
A developer explains how the Model Context Protocol (MCP) differs from traditional APIs by enabling ongoing, context-aware conversations between systems rather than one-off requests. MCP allows programs to remember past …
Anthropic's Claude Sonnet 5 reduces costs for agentic AI, while agent frameworks like Konsistent, RF-DETR Keypoint, Search Toolkit, and Koog 1.0 ship stable APIs and tooling. Konsistent enforces structural code conventio…
Neander, a purpose-built language for agentic systems, makes API discovery a first-class language feature through its 'discover' verb. The language unifies finding and calling APIs into a single 'submitProgram' operation…
SGLang development is being augmented with agent-assisted workflows that encode procedural engineering knowledge into executable skills, covering LLM serving, GPU kernels, diffusion pipelines, and production incident han…
Panoptes, an AI audit and alignment layer, launches to provide universal observability, audit trails, and policy enforcement for AI coding agents like Claude Code, OpenAI Codex, Google Antigravity, and Hermes Agent. The …
YouTuber Husk IRL placed three phones running different conversational AI agents together and asked them to count to 100. The agents negotiated a shared approach but repeatedly agreed with and reinforced each other, fail…
Sysdig's threat research team caught what appears to be the first ransomware attack run almost entirely by a large language model, named JADEPUFFER. The AI agent broke into a server, stole credentials, encrypted a databa…
A developer built storygen-skill, a coding agent that uses Pixar's Braintrust process to write and revise illustrated stories, producing two sample books. The tool applies structured self-review with AI personas acting a…
Stitch, a free open-source desktop app, launches to manage emails, record meetings, track tasks, and automate workflows locally without cloud dependency. The app runs on macOS and Windows, supports multiple AI providers,…
A developer argues that Git is more essential than ever in the age of AI-generated code. The developer explains that Git commits provide safe rollback points for AI changes, diffs are crucial for verifying AI output, and…
A blog post on bawes.net describes the integration of spatial computing with the Model Context Protocol (MCP) in an open virtual-world platform called Universe, where autonomous agents use MCP to access databases, APIs, …
A developer gave an AI agent named Hermes its own email address using AgentMail, a service designed for AI agents. The agent can now receive emails, sign up for services, and subscribe to newsletters, effectively gaining…
A developer has created a 10-line prompt that transforms ChatGPT into a fully autonomous AI agent capable of planning, executing, evaluating, and improving tasks without human intervention. The prompt leverages the ReAct…
AI programming agents like Claude Code, Cursor, and Windsurf run on an AgentLoop analogous to a CPU's fetch-decode-execute cycle, but running them locally on laptops causes severe compute bottlenecks and session instabil…
Artificial Analysis launched AutomationBench-AA, an independent leaderboard for Zapier's AutomationBench, testing AI agents on 657 real SaaS workflow automation tasks across 40 simulated apps. Claude Fable 5 leads with 4…