A lot of people seem convinced that the point of AI coding is to write low-quality code as fast as possible. Spew out barely-passable slop, open massive PRs, and merge them unvetted. Ship it! But the thing is, LLMs are very flexible. And you can use them just as effectively to write high-quality code more…

Anthropic introduced 28 security and compliance tool integrations to help IT and security teams govern Claude.

An early update on what we've learned from Project Glasswing.

In recent weeks, we pointed Mythos and other security-focused LLMs at live code across critical parts of our infrastructure. We share what we observed, the models’ strengths and weaknesses, and what the work around them needs to look like before any of it can scale.

We're launching Claude for Small Business, a package of connectors and ready-to-run workflows that put Claude inside the tools small businesses use every day.

Find the local LLM that actually runs and performs best on your hardware. Ranked by real, recency-aware benchmarks, not parameter count. One command, run it instantly. - Andyyyy64/whichllm

Local AI models should be the default.

Experiments with getting usable outputs out of local models on a standard Macbook

Demystifying evals for AI agents

AI-generated images have left us questioning what is real. But the godfather of digital forensics, Hany Farid, is not giving up

An encyclopedia of a universe that does not exist until you visit it.

What happened? In 1.110, we added a setting to add Copilot as coauthor in commit messages by appending Co-authored-by: Copilot [email protected]. The setting git.addAICoAuthor has three different values: off - no attribution no matter w...

Use Claude Code's autonomous agent loop with DeepSeek V4 Pro, OpenRouter, or any Anthropic-compatible backend. Same UX, 17x cheaper. - aattaran/deepclaude

Visual Studio Code. Contribute to microsoft/vscode development by creating an account on GitHub.

On April 21, 2026, a major breakthrough in cybersecurity happened: leading standardization initiatives gathered in Washington DC and agreed to begin coordinating collectively on AI security. A personal dream come true. The result: MOSAIC: Multi-Organization Secure AI Coordination. The goal: turn a fragmented landscape into clear, consistent standards and guidelines, to deal with the mounting risks of AI.

This important step was taken at the AI Security Policy Forum, organised and led by the OWASP AI Exchange, with SANS Institute as co-host - convening standard makers and policy stakeholders.

The initiatives at the table included: 👉 BIML (Berryville Institute of Machine Learning) 👉 Center for Internet Security (CIS) 👉 Cloud Security Alliance (CSA) 👉 Coalition for Secure AI (CoSAI) 👉 National Institute of Standards and Technology (NIST) 👉 OWASP AI Exchange (AIX) 👉 OWASP GenAI Security Project 👉 SANS Institute

The group agreed that it is now more important than ever to coordinate around the rapidly evolving possibilities and challenges of AI, as AI security risks mount.

One of the next steps is to provide a standardized map of the participating initiatives and a communication platform to exchange insights on a first list of identified topics (e.g., aligning with other initiatives such as SC42, building on OpenCRE, consensus on definitions), improve consistency, clarity, quality, and prevent unnecessary duplication. The idea is to move fast while maintaining independence and with lightweight coordination - not add more committees.

In addition to the organizations mentioned, the discussion also included journalists, representatives from International Telecommunication Union (ITU), The Aspen Institute, academia, and government — providing valuable perspectives on developments in both policy and industry. This helped prioritize the topics to focus on.

In the picture, from left to right, standing to sitting: Disesdi Shoshana Cox (AIX), Gary McGraw(BIML), Rob van der Veer (AIX), Anonymous, Duncan Sparrell, John Yeoh (CSA), Rock Lambros (GenAI), Norma Krayem, Brian Calkin (CIS), Matt Altomare (Aspen), Omar Santos (CoSAI), Aruneesh Salhotra (AIX), Jonathan Gibson (The Dispatch), Apostol Vassilev (NIST), Rhea Nygard, Ken Huang, Lav Varshney (Stony Brook University), Sounil Yu, and Sharon Goldman (Fortune)

Not in the picture, but involved, in alphabetical order: Rob T. Lee (SANS), Ryan Galluzzo (NIST), Soribel F.

A big thank you to: 👏 Disesdi Shoshana Cox for her idea to bring everybody together in a room to fulfil the connecting mission of the Exchange 👏 The amazing thinktank at the AI Exchange 👏 Spyros Gasteratos for his work on OpenCRE 👏 Violeta Klein, CISSP, CEFA for shaping the story for the Forum 👏 Straiker, Casco (YC X25), AI Security Academy, and SANS for supporting the Forum. 👏 Software Improvement Group for donating the original threat model and initiating the AI Exchange

Let’s make AI a success! | 28 comments on LinkedIn

A TDD-driven iterative feedback loop for software development. 16 cohesive Claude Code skills walk an idea from brainstorm → plan → execute → iterate, with checkpoints throughout. - evanklem/evanflow

Multi-lens code audit tool — 280 expert AI agents for code review, security testing, and infrastructure auditing - TheMorpheus407/RepoLens

Is security spending more tokens than your attacker?

Our latest model, Claude Opus 4.7, is now generally available. Opus 4.7 is a notable improvement on Opus 4.6 in advanced software engineering, with particular gains on the most difficult tasks.

A demo combining LeCroy oscilloscope control, SPICE simulation, and Claude Code.