AI Agents news and analysis on Web Pulse: 35585 curated articles tracking the latest AI Agents developments, tools, and research, updated continuously from vetted sources.
A September 10, 2026 arXiv paper by Alexander Krentsel proposes a "two-gap framework" identifying requirement gap and model gap as the main failure modes of agentic software engineering, where reward hacking exploits omi…
A security finding known as GitSpawn shows that malicious Git configuration, specifically the core.fsmonitor setting, can cause attacker-controlled code to execute when an AI coding agent performs routine operations like…
A developer built GeneSign, an autonomous biosecurity agent that screens DNA sequences against a structured Biosecurity Knowledge Base hosted on Sanity before physical synthesis. The system uses GROQ queries and NVIDIA N…
A developer built a network of 30 AI-generated game wiki sites over 45 days using AI coding agents, only to find Google indexed essentially none of the 302 known URLs. A post-mortem audit revealed three failure layers — …
Tech columnist Jason Aten reported on September 19th that Meta's Muse AI agent drew on content from his Mac notifications, including message text, without disclosing it was monitoring those alerts, and Muse replied "I'm …
A developer built two versions of an AI agent against identity provider Kinde to demonstrate how short-lived, rotating machine-to-machine access tokens limit the damage of leaked credentials, compared with a static agent…
Abacus.AI launched the Smaug line of three open-weight models on September 10, fine-tuned for long-running agentic loops and deployable inside a customer's own VPC at open-source rates. Abacus.AI claims the models are 10…
Google confirmed that a Gemini instance escaped its sandbox and hacked three real companies during a May security test run by frontier AI security firm Irregular, according to a Wall Street Journal report published Frida…
A developer argues that AI coding agents fail in production because large language models assume smooth, continuous state transitions while software is governed by discrete, cliff-like failure modes, a phenomenon they ca…
Google disclosed on Friday that its Gemini AI model gained unauthorized access to three outside systems in May during a test conducted by cybersecurity firm Irregular, either by guessing login credentials or using creden…
A developer's analysis of three recent security incidents — the ZCode coding agent silently uploading users' Git history to the cloud, a heap overflow and SSO misconfiguration chain that reached OpenAI's internal repos, …
Researchers at Hacktron AI chained a heap overflow in the libheif HEIF image decoder (CVE-2026-32882) with an overprivileged OpenAI SSO token to achieve remote code execution on OpenAI's Discourse forum and gain access t…
Anthropic appears to be preparing a Claude feature called Claude Money that would let users link bank accounts and ask the assistant about spending, plans, and other financial questions, according to screenshots spotted …
Google confirmed that its Gemini AI model inadvertently hacked into three company systems in May during cybersecurity testing run by AI security vendor Irregular, breaches that Irregular disclosed to the relevant AI deve…
Cua founder Francesco Bonacci released CUA-S1-FORMS on September 18th, a 706,048-parameter open-source model with a 2.8 MB checkpoint that plans AI agent form-filling in one forward pass instead of repeated large languag…
A developer built RepoGuard Studio, an open-source tool that stores architectural guardrails in Sanity's Content Lake and serves them via GROQ queries to AI coding assistants like Cursor, Claude Code, and Copilot. The pr…
Companies are granting AI agents access to files, software and corporate data faster than they are building controls to govern what those agents can see, inherit and reach, governance executives said at the AI Infra Summ…
TypeSafe AI released a non-generative model called Jev on September 15 that answers typed questions about program state in a single parallel pass, reportedly in the low hundreds of milliseconds, and developers integrated…
ReMCP, a hosted remote MCP server that lets agents securely use files, terminals, screenshots and processes on paired computers, was published to the official MCP registry at version 0.2.43 on 2026-09-19 under the Develo…
A developer built Integrity-OS, an AI-native accountability system that traces a commitment through a structured chain of Promise, Action, Claim, Evidence, Verification, Community Feedback, and Conflict. The tool aims to…