AI Agents news and analysis on Web Pulse: 49340 curated articles tracking the latest AI Agents developments, tools, and research, updated continuously from vetted sources.
A practitioner running turnstone LLM agent instances reported that removing all local tools and execution capability so that MCP is the only path out of the workers has been only partly effective at containment. The oper…
OpenAI published Red Card, a webcam-powered Codex skill hosted at github.com/openai/redcard that lets users leave a Google Meet call by holding a physical red card up to their Mac's camera. On detection, Red Card sends a…
A developer built a retrieval-augmented generation service with hybrid dense+BM25 retrieval, a 0.15 relevance threshold meant to trigger refusal, and an extractive mock generator that quotes cited chunks, then evaluated …
IBM published a security bulletin on October 28, updated October 29, disclosing four vulnerabilities in Langflow, the open-source tool for building AI agents and workflows, including two critical flaws scored 9.8 that re…
A GitHub project called roast-room lets users convene a council of AI advisors inside Anthropic's Claude Code to debate a business decision and return a scored verdict plus a downloadable PDF. The tool ships with 12 advi…
A developer running a daily autonomous Claude Code agent on a Windows PC found that a series-grouping update returned an HTTP 500 even though the change had already been applied server-side, causing the scheduled task to…
Meta's Muse personal AI agent has paid roughly $1.2 million in bug bounties, including $300,000 for an authentication flaw found four months before launch, David Singleton said in a thread on X and an accompanying blog p…
An engineer published an architectural blueprint for an automated pre-production evaluation harness that uses Amazon Bedrock, telemetry tracing, and LLM-as-a-Judge quality gates to catch AI contact center agents that vio…
Mercor routed 18 million requests and 63 billion tokens through Vercel AI Gateway in its first 10 days on the platform without hitting rate limits, cutting time-sensitive batch jobs from more than 12 hours to 2 to 3 hour…
Zapier's Powered by Zapier team now builds fully interactive, multi-view partner prototypes in under two hours using Vercel's v0, replacing a process that previously took weeks of screenshots, design time and frontend re…
Picos is a new agent harness that gives large language models a custom 32-bit CPU called smol32 with only 32 instructions (0x00–0x1f), 16 registers (r0–r15), a single flat RAM array, raw disk storage, and a 256 × 192 RGB…
A developer published an architectural guide for building production-grade Model Context Protocol (MCP) servers in pure Python using asyncio, json, and Pydantic V2. The guide covers JSON-RPC 2.0 wire format, Pydantic dat…
A developer outlined a pattern for preventing stale AI agent workers from repeating side effects by combining a durable run lease with a monotonically increasing fencing token. The lease, stored in a database or queue wi…
A developer built two open-source projects during Hacktoberfest 2026: SlothMode, a local-first AI productivity companion that enforces scheduled breaks using open-weight models and a local .ics calendar file, and lux, a …
Researchers introduced MAServeBench, a benchmark for evaluating multi-agent LLM serving systems under hierarchical collaboration workloads, addressing how interdependent model and tool operations become ready at differen…
Agent Remote v0.8.8, an MIT-licensed open-source tool, lets users run Claude Code or Codex sessions on a Mac and follow, reply to, and finish the same live session from an iPhone or another Mac, pairing devices by QR cod…
Simon Willison released llm 0.37 on 2 November, adding the gpt-6.1-sol and gpt-6.1-terra models, an `llm openai endpoint add` command for arbitrary OpenAI-compatible endpoints, and a switch of the default model to gpt-6-…
An engineer outlined a safety checklist for teams connecting always-on coding agents to GitHub repositories, drawing on the public dsh-superagent research prototype hosted under the deepseek-ai organization, which its RE…
PhoneUse launched a tool that gives AI agents control of real retail iPhones and Android devices, exposing 12 tools including phone.acquire, screen.read, tap, type, swipe, app.open and sms.read through an MCP server or A…
A field engineer reports that replacing six separate notations (Makefiles, Gherkin, EBNF, init.d scripts, and others) with a single Smalltalk-80 keyword-message grammar produced a 100% pass rate (120/120) in a 360-trial …