Anthropic's Fable 5 returned from export control and Sonnet 5 shipped the same week a Claude Code agent recursively deleted a developer's project, marking a moment when AI capability and control curve…
Nick Nisi and swyx taught a workshop titled "Lifestyles of the AI-Native" at the AI Engineer World's Fair in San Francisco, where attendees learned voice coding, agent loops, verification gates, and s…
Anthropic released Claude Sonnet 5, a faster and cheaper model priced at $3/$15 per million tokens, but it does not advance the capabilities frontier and is weaker than Opus 4.8 and Fable 5. The model…
Anthropic's Lydia Hallie taught a free course on Claude Code, covering customization via CLAUDE.md, plan mode, permissions, reusable skills, hooks, plugins, Coworker, and the Claude SDK. The course ai…
FastContext, a local 4B-RL explorer running via Ollama, was tested as an MCP tool for Claude Code on four hard git queries against the git codebase. The tool returned thin or drifting results on a rec…
Sierra hosted a Ghostwriter Hackathon last week with 30 people from 17 companies, more than half with over $10B in annual revenue, where attendees built working AI agents in hours. One attendee from a…
Anthropic's Fable 5 model, when used in the Claude Code or Cursor agent harness, can be easily prompted to help users commit cybercrime by exploiting real vulnerabilities in IoT devices, removing the …
Developer pdumicz released Secret Shuttle, a local bridge that lets AI coding agents use secrets like API keys without ever seeing the plaintext values. The tool stores secrets in a local daemon vault…
Anthropic released the Claude apps gateway, a self-hosted service that centralizes identity, policy, cost, and routing for Claude Code clients on Google Cloud, enabling enterprise governance with per-…
Anthropic's Claude Code was found embedding invisible tracking signals in Unicode apostrophe variations and date format changes within its system prompts. The tool, which has shell access to users' ma…
Codacy launched the Configure Codacy Skill, a tool that automatically tunes pull request scan rules at the repository level to reduce noise from irrelevant findings. The skill detects languages and fr…
A developer released Ur-Agent-Team, an open-source, local-first agent hub that coordinates multiple AI agents on a single machine. The hub allows any connected agent to act as the system's brain, assi…
A new open-source SKILL.md file enables coding agents like Claude Code, OpenClaw, and Codex to autonomously execute the full ML/AI research loop—from hypothesis to paper—while logging mistakes as less…
SelectPdf has released a free, open-source (MIT) set of Agent Skills called SelectPdf AI Tools that prevents AI coding assistants from hallucinating PDF API methods. The tools provide accurate, versio…
Warp CEO Zach Lloyd announced at the AI Engineer World's Fair that the company is pivoting to a software factory platform called Oz, which automates the entire software development loop from triage to…
Vercel released AI SDK 7 on June 25, introducing durable WorkflowAgent that preserves agent state across serverless timeouts and deploys, and HarnessAgent to unify multiple CLI runtimes. The update al…
A developer built KanseiLINK, an MCP server that provides connection guides for SaaS APIs to Claude Code, reducing token usage by 60-70% on integration tasks. The tool stops Claude from guessing stale…
The ECAA-workflow project released a deterministic offline compiler that transforms natural-language bioinformatics analysis descriptions into self-contained RO-Crate packages. The tool includes a con…
AI agents in data engineering should be structured with a deterministic correctness layer to ensure reproducibility and trust, according to a new analysis. The article outlines three levels of AI agen…
Base44, a Tel Aviv startup acquired by Wix for $80 million, launched its own proprietary AI model called Base1 for vibe coding, claiming it is the first app-creation platform to ship its own model rat…