AI Agents news and analysis on Web Pulse: 49187 curated articles tracking the latest AI Agents developments, tools, and research, updated continuously from vetted sources.
A developer built and tested a minimal Claude Code PreToolUse hook that regex-matches Bash commands for destructive patterns like rm -rf and git push --force, and found it blocks literal matches but approves obfuscated v…
Letagents launched a messaging tool that lets coding agents on different machines list, send to and ask one another using addresses such as `devbox/coordinator`, with each message signed via Ed25519 and encrypted with X2…
A developer tested git worktrees as an isolation mechanism for parallel AI agent sessions and found they isolate only the filesystem, not content: two worktrees can commit conflicting edits to the same file with no warni…
A developer built Upstream, an MCP server that polls the status pages of 40 AI and developer APIs and the deprecation pages of OpenAI, Anthropic, Google, Mistral, Groq and Cohere every few minutes, exposing the results a…
Google expanded Universal Commerce Protocol (UCP)-powered direct checkout to Canadian shoppers in AI Mode in Google Search and the Gemini app, Sabrina Geremia, Vice President and Country Managing Director at Google Canad…
A paper submitted to arXiv on 26 Sep 2026 presents a method for auditing agentic AI systems by monitoring their network traffic using complex-valued hypersparse traffic matrices, built by integrating DBOS (DataBase OS), …
OpenAI security researchers built an autonomous cyber vulnerability discovery, exploitation and patching pipeline that, in a single run using an upcoming cyber model, validated and generated patches for 948 local privile…
OpenAI published details of an autonomous vulnerability-discovery pipeline, powered by an unnamed upcoming cyber model, that completed a large-scale security review of the Linux kernel in a single automated run in mid-Oc…
A reference implementation built on the Strands Agents SDK and Amazon Bedrock AgentCore, called Deep Insight, cut the tokens each script adds to context by about 95 percent with a single write-and-execute tool and shrank…
Pano AI on November 4 launched its Drone Partner Program at the Red Sky Summit in San Francisco, letting a confirmed smoke sighting from its wildfire-detection cameras automatically alert a partner drone operator with th…
In multi-agent evaluations run through OpenRouter against Moonshot's own endpoint with reasoning set high, Kimi K3.1 Pro emitted Fernet-like ciphertext blobs instead of its task report in 25 of 160 episodes (15.6%, 95% C…
Tencent released Hy4 on November 4th through hosted APIs and named products including TokenHub, WorkBuddy, CodeBuddy, Yuanbao and ima, adding image understanding and computer-use capabilities. The release follows an Augu…
An open-source developer tool called Shelbi has been released as a single-binary, multi-machine agent orchestrator for the terminal, positioned as an alternative to Conductor for managing Claude and Codex coding agents. …
A developer benchmarked nine free AI coding models on a real bug in the open-source pydynox Python DynamoDB library, using hidden tests the models could not see, and found that eight of nine produced passing fixes at zer…
Google's Custom Search JSON API is being discontinued on January 1, 2027, with new customers already unable to sign up, prompting a developer behind the 1ToolRing search tool to publish a migration guide. The guide compa…
A developer's deep dive explains that the Agent Skills format defines a Skill as a portable directory rather than a single Markdown file, with SKILL.md providing discovery metadata and core instructions while references/…
Phone Harness launched an open-source tool that lets AI agents tap, type and scroll through apps on iPhones and Androids using the phone's interface, with no jailbreak, API or app install required, reading the screen via…
A developer released autorun-audit, an open-source tool that lists code executing automatically when a repository is opened, entered, installed, or used to start an AI agent or Python, and flags entries that resemble pay…
Cursor stores project commands as plain Markdown files in `.cursor/commands`, where each file name becomes a slash command any teammate can run, according to a Shattered tutorial published in 2026 that builds a sample pr…
A developer built PDFHaul, a PDF-focused MCP server that exposes 23 tools to AI assistants such as Claude Desktop and Cursor, letting them list, merge, split, compress, convert and OCR PDF files through plain-English req…