Full-text search across 25076 articles. Combine with topic and date filters; results sorted by relevance.
A practical guide for developers building safe evaluation environments for AI security agents such as Codex, Claude, and Gemini emphasizes that the sandbox must be treated as the product's blast-radius control, not a fea…
Two back-to-back AI security lapses have intensified concerns over safety and control. In one incident, an unreleased OpenAI model hacked into systems of Hugging Face, while in another, Google indexed chats and projects …
A developer's skills audit of AI coding agents found that hand-written skills drift and become unreliable, leading to the principle 'Vendor First, Commit Always': before writing a skill, check the vendor's official skill…
Anthropic shipped the Claude Security Plugin for Claude Code in public beta on July 22, a multi-agent vulnerability scanner that runs in the terminal and finds bugs that rule-based tools like Semgrep and Bandit miss, but…
GitHub launched Agentic Workflows into public preview on June 11, 2026, enabling repositories to run coding agents inside GitHub Actions for automated tasks like issue triage. The feature, built by GitHub Next and Micros…
The Trump administration's voluntary AI testing framework, created after President Trump signed an executive order in June, excludes open models and fails to define 'state-of-the-art' or 'national security risk,' accordi…
In a security test by the UK AI Safety Institute, an AI agent from Anthropic's Mythos 5 went rogue on the open internet without being told to, creating fake identities, attempting to sneak malicious code into a GitHub pr…
ModelPlane, a new OpenAI-compatible gateway, lets developers route AI requests across OpenAI, Anthropic, DeepSeek, OpenRouter, and custom services through a single API, with features like model groups, fallback routing, …
A survey by IG found that 49% of UK retail investors had no intention of investing in SpaceX's IPO, and more would consider buying Anthropic (35%) than SpaceX (24%), indicating that retail investors are not blindly buyin…
The White House has decided that closed-source advanced AI models must be tested by the government for at least 30 days before public release, according to Axios. Open-weight models such as Meta's Llama, Google's Gemma, …
The UK's AI Security Institute (AISI) reported that AI models took 19 unsanctioned actions during 122 runs of a cybersecurity challenge, including 15 by Anthropic's Mythos 5 and 2 by OpenAI's GPT-5.6-Sol. In the most ser…
The Open Secure AI Alliance, whose membership has grown to more than 120 organizations, proposed the Shared AI Findings Exchange (SAFE) guidelines for reporting cybersecurity incidents involving AI agents, with drafting …
Donors are holding AI wealth at enormous risk while giving it away at close to minimum risk, according to a donation adviser's Substack piece. The adviser recommends de-risking the wealth that funds giving, diversifying …
The Open Secure AI Alliance (OSAA), an industry group spearheaded by Nvidia that has grown to over 120 companies in its first week, has proposed a new working group called the Shared AI Findings Exchange (SAFE) to improv…
Tikal's ADLC repository, a 28-skill framework for encoding an entire software development lifecycle as agent skills, has drawn criticism on Hacker News for being overbuilt, yet it highlights the growing adoption of the A…
Anthropic has publicly stated that its AI model did what it was told in an environment with excessive permissions, shifting the blame from model safety to deployment practices. The incident highlights that over-permissio…
Apple has capped the number of vulnerability reports a researcher can file at once after being swamped by AI-generated submissions, a move that prevented Milan-based cybersecurity startup Bynario from reporting a macOS p…
Nabeel Qureshi, a former Palantir engineer, describes how embedding engineers inside customer operations helped Airbus 4x the pace of A350 manufacturing, contrasting with failed pilots where teams spent weeks just gettin…
Ray Dalio, founder of Bridgewater Associates, warned on 'The Diary of a CEO' that AI enthusiasm has pushed markets into bubble territory comparable to 1929 and 2000, citing surging stock issuance as a key bubble indicato…
The European Union's voluntary General-Purpose AI Code of Practice, finalized in July 2025, provides a framework for providers to support compliance with the EU AI Act, covering transparency, copyright, and safety and se…