AI Agents news and analysis on Web Pulse: 49320 curated articles tracking the latest AI Agents developments, tools, and research, updated continuously from vetted sources.
Claude Opus 5.5 leads the Agent Index v1.2 preview, a November 2, 2026 snapshot of seven published agent benchmarks, with a composite score of 90.4 out of 100, ahead of GPT-6 Astra at 87.6 and Claude Sonnet 5.5 at 85.8. …
NVIDIA released Auto-Ontology, an Apache 2.0 open context layer that compiles database metadata into business terms, reviewed SQL measures and table relationships and serves it through MCP, agent skills and a REST API, d…
Developer Chris Smith disclosed GitHub advisory GHSA-87xc-jgvf-hmvg, a server-side request forgery (CWE-918) in his reverse proxy Centauri that carries a CVSS score of 10.0, after Anthropic's Claude Code flagged the flaw…
GitHits CLI 0.38.1, released November 3, 2026, lets coding agents resolve a package version to an exact source revision and read dependency code, as shown in a walkthrough where an agent looked up httpx 0.28.1 and found …
An analysis of 51,000 coding-agent sessions from 190 engineers found that top-quartile users get 7.8 minutes of agent work per minute of their own time versus about half a minute for the bottom quartile — a 15x return on…
Craft CMS 5.12.0 added an authenticated HTTP MCP server that MCP clients can connect to after a user signs in and authorizes them via OAuth, with clients acting as that user within their permissions and a granted access …
An external audit of 61 official remote MCP servers requiring login found that all implement HTTP 401 challenges and PKCE with S256, but adoption of newer parts of the 2026-07-28 MCP authorization spec lags: only 25 adve…
A solo developer in Japan documented how the Claude desktop app's scheduled tasks silently stopped firing for days in September 2026, with the app logging 4,322 "Deferring check: net.isOnline() is false" lines over 72 ho…
A paper submitted on 2 Nov 2026 introduces TLAPS-Bench, a benchmark and execution harness for building machine-checked proofs of real-world system specifications in TLA+ using agentic AI atop the TLA+ Proof System (TLAPS…
Developer Ibrahim Kobeissy released Harneo v0.2, an MIT-licensed open-source harness manager that keeps Claude Code skills, commands, subagents, hooks, settings.json, MCP servers and CLAUDE.md files in one library outsid…
A 16-year-old developer has launched Dermioz, a private-beta enterprise AI memory tool that tracks who said each fact and when, resolves conflicting sources, and flags outdated information, claiming it retained a changed…
Cloudy launched a free AI chat service and a terminal coding agent called Cloudy Code that share one account, offering 6 switchable models, 25 messages and 70 agent credits per day per account, and a stated policy that c…
Flox released a GitHub repository, agentic-development-with-flox, that consolidates nearly 80 turnkey Flox environments for agentic development, covering harnesses including Claude Code, Codex, OpenCode, Hermes Agent, Pi…
A developer recommends Grafana with Prometheus as the strongest free starting point for AI agent metrics dashboards, arguing teams should track four per-run measurements — total loop duration, model-call duration, retry …
A developer built LurkAPI, a social media data API that returns public posts, profiles, comments, video transcripts and ad-library data from 16 platforms as JSON via a single GET request, across 88 endpoints. The same en…
A developer outlines a minimal, low-cost metrics dashboard approach for small teams running AI agent loops, arguing that signal quality rather than license cost should drive the choice. The recommended schema records one…
A developer behind Linkmortem, a dead-link checker that uses the open-weights Gemma 4 model to classify pages code cannot judge, reported that an AI crawler-targeting page attempted an indirect prompt injection with the …
A security report filed against Postman's desktop app found that on launch it automatically wrote Codex and Claude Code agent files, installed a startup hook instructing the coding agent to load Postman's skill before no…
Revefi released a Claude Code plugin, installable via `/plugin marketplace add revefi/agent-analyzer`, that scans the last 14 days or 100 most recent sessions (capped at 200 MB compressed), redacts them locally, and—only…
A developer built Shield, a human-in-the-loop data quality agent, during the Flip the Switch hackathon on Switch 0.31.0, running it in a GitHub Codespace with fictional data. The agent checks a daily table, reports findi…