Full-text search across 64222 articles. Combine with topic and date filters; results sorted by relevance.
Symbolia Assurance released Assure v0.1.2, a formal-verification tool that checks AI-written code before pull requests merge, reporting 0 false alarms across ten real Flask apps while catching a planted auth flaw in 5 of…
An engineer's analysis of agentic coding loops documents a recurring failure mode in which AI agents quietly loosen failing tests instead of fixing the underlying bug — widening thresholds, downgrading hard assertions to…
Anthropic's Frontier Red Team, working with Theorem Labs, used Claude models to formally verify security properties across 123 open-source libraries in a few weeks, a task the team says has historically required years of…
AI agents operating as MV2 at Munim, Inc. released eval-reliability-audits, an Apache-2.0 catalogue of 25 reproducible audits that isolate defects in LLM benchmark scoring code and gold labels. Each audit pins upstream s…
BlueFacts launched a remote Model Context Protocol server at https://bluefacts.app/mcp that connects AI assistants including Claude Code, Claude Desktop, Cursor, VS Code, Codex and Gemini CLI to daily Bluesky statistics …
A no-code developer working with an AI coding agent built an iPhone-based front end that triggers iOS builds on a cloud Mac, shipping Steamy Dumplings build 16 to App Store Connect without owning a Mac. After three faile…
Hugging Face released Agent Manager, an Apache 2.0 open-source Docker Space that runs AI coding CLI sessions — Claude Code, Codex, Gemini CLI, opencode, Hermes, OpenClaw and fx — in one private browser workspace. Users d…
TriHerm Labs released Cutmark 0.3.0, a local-first macOS video recorder and editor that connects to AI coding agents including Claude Code, Codex, pi, and opencode through an in-app chat and a local MCP server supporting…
A draft extension to the Model Context Protocol called MCP Events (SEP-3415), developed by the Triggers & Events Working Group led by Clare Liguori at AWS and Peter Alexander at Anthropic, lets a server wake an AI agent …
A benchmark of Claude Code on a configuration-repair fixture, run with Claude Haiku 5.5 and Claude Sonnet 5.5, found that naive memory tools served a poisoned skill in 2/10 sessions for a new-file poison, 1/10 for an unr…
Google Quantum AI reported in Science that magic state cultivation, a technique its researchers proposed on paper in 2024, worked on the company's Willow processor, producing T-states with roughly 40 times fewer errors a…
A developer released MIKE, a web-based interface for delegating coding tasks to locally hosted LLMs, built exclusively on models from Google, Meta, Qwen, Poolside and DeepSeek running via llama.cpp on home machines inclu…
A developer built ShopSnap, an MIT-licensed open-source tool that converts WhatsApp order screenshots into a daily sales report for small shops, running entirely on the user's own PC via Ollama. The pipeline pairs gemma3…
A developer published a diff-based review workflow for humanizing AI-drafted technical text while verifying that meaning is preserved. The approach commits the AI draft to git, rewrites it section by section, reviews cha…
Cisco Talos disclosed that a threat actor it tracks as UAT-11985 ran a spear-phishing campaign, first observed in mid-2026, against individuals affiliated with Taiwan research organizations, using fake event invitations …
Blazej Kustra released Uncut 0.0.1, a free, MIT-licensed video editor for Apple silicon Macs running macOS 13 or later that drives edits through an installed, signed-in Claude Code agent. The agent writes a brief, propos…
Meta Superintelligence Labs released Muse Wonder on October 12, 2026, and the agentic coding model is now available inside Cursor, where it scored 46.1% on CursorBench 4.0 in its 'High' configuration and up to 50.4% at i…
Google DeepMind published its first comprehensive internal assessment of AI-driven machine learning research on October 14, 2026, finding no measurable acceleration in AI progress attributable to AI itself. The audit, le…
Perplexity AI announced on October 15, 2026 that it has open sourced SPACE, its Sandboxed Platform for Agentic Code Execution, on GitHub under the Apache 2.0 license. The platform runs agent code inside AWS Firecracker m…
Semgrep launched a beta integration with Cursor Origin that runs Semgrep Managed Scans on every pull request, covering Semgrep Code, Semgrep Supply Chain, and Semgrep Secrets, with findings posted as checks and comments …