AI Agents news and analysis on Web Pulse: 49353 curated articles tracking the latest AI Agents developments, tools, and research, updated continuously from vetted sources.
Softcane released an alpha Kubernetes operator, codex-session-operator, that runs OpenAI coding agents against Git repositories inside a cluster, with OpenAI handling the agent loop and model inference while Kubernetes A…
A guide on evaluating AI agents recommends measuring six dimensions — task success, trajectory, policy and safety, reliability, efficiency, and interaction — at three levels (single step, whole task, whole conversation),…
A new paper circulating on X reports that newer AI models help peer agents more often than older systems, even when doing so lowers their own scores, a finding discussed alongside OpenAI's GPT-5.6 Sol, which launched pub…
Developer Damir Sokolovsky released Unslop, a GitHub-hosted skill for Claude Code and Codex that audits "vibecoded, AI-looking" interfaces and proposes a design direction before editing. The skill flags patterns such as …
A study posted to arXiv on 25 October 2026 by Hui Sun and co-authors used Claude Code to migrate ten of 148 PyTorch ML research projects from NVIDIA GPUs to Google Cloud TPUs, finding that 53 of 70 execution attempts fai…
Anthropic released message threads in public beta on the Claude Platform, letting agents send only new messages and the previous response ID each turn instead of resending full conversation history. With message threads …
A developer published a Claude Code skill that reviews a GitHub pull request by number and returns an approve or reject verdict. The skill runs a code-review pass, rejects only for real defects such as bugs, data loss, s…
An autonomous AI engineer operating as Blucca built Goalkeeper, a small JavaScript controller for the battle.training arena that drafts four APCs and holds weighted objectives, winning all 24 local house-bot games while …
LiteLLM, an open-source library from BerriAI, lets developers call more than 100 LLM providers through a single OpenAI-format completion call, so switching between OpenAI's gpt-4o and Anthropic's claude-sonnet-4-20250514…
Vercel MCP now accepts connections from any MCP client that publishes a Client ID Metadata Document (CIMD), a JSON file on its own domain whose HTTPS URL serves as the OAuth client_id, eliminating Vercel's prior manual r…
Developer Addy Osmani released agent-skills, a set of 25 skills for Claude Code, Cursor and Codex that installs via a single npx command and forces coding agents through a spec-plan-build-test workflow. The skills add /s…
A developer built the Blupe playground, an open-source interface that lets users control real robot arms with AI models, submit task prompts, and watch live video of runs, with code published on GitHub under the blupe-re…
A developer automated a niche technical newsletter and code-auditing side project using Python and the Anthropic API, running the pipeline on a cron job to generate subscriber reports and process micro-payments without m…
A CloudHike developer published a walkthrough for building an AI summarizer that offloads slow model calls to background jobs with retries, signed webhooks and duplicate protection, all provisioned through a single MCP c…
A cofounder of an in-progress AI integration lab released Switchboard, an AMQP 0.9.1 message broker written in Rust that an LLM implemented from the protocol specification plus requirements for clustered Raft and support…
Sierra redesigned its Agent Studio around Ghostwriter, its AI agent for building agents, so that investigating a problem, editing a journey, running simulations and testing a fix now happen in a single conversation rathe…
A Momentic engineering guide recommends that reviewers test coding-agent pull requests against each PR's deployed Vercel or Netlify preview URL before reviewing the diff, because code review alone cannot show whether cha…
A new guide details how teams can run a pre-release AI bug bash by giving an AI QA agent called Mo a written brief and a reachable staging or preview URL, without writing test files beforehand. The brief should name the …
WorkOS added Fireworks AI as a bring-your-own-key inference provider for Airlock, its governing agent that reviews proposed AI agent actions against team policies, the company said in an early-access announcement. Admini…
A developer built an MCP server for Apparent, a self-built mobile app analytics tool that reads App Store Connect, Google Play, RevenueCat and its own SDK, so that Claude Code can answer questions like "how are my apps d…