Full-text search across 26394 articles. Combine with topic and date filters; results sorted by relevance.
A developer's blog post documents running Claude Code and Codex coding agents inside Apple's open-source `container` tool, which starts a separate lightweight Linux VM with its own kernel for every container, unlike Dock…
OpenMausBot, an independent open-source project formerly called OpenGrokBot, released a local-first alternative to xAI's Grok Bot under the Apache 2.0 license, running bots on the claude, codex and grok CLIs already inst…
Cisco open sourced Project Quarry, a neurosymbolic detection rule language and evaluation engine for identifying multi-step attacks and misaligned behavior in AI agents, published on GitHub under cisco-ai-defense. Cisco …
Developer Adam K. Dean released claude-chalk, a Claude Code plugin that lets Claude color its own terminal replies using chalk-template markup syntax, installed via `/plugin install chalk --marketplace adamkdean/claude-c…
Cisco has open sourced a tool for detecting AI agents conducting hacking sprees, The Register reported. The release gives defenders a way to identify autonomous agent activity during attacks, a capability that matters as…
A developer built The Ofrenda Witness, a ten-minute browser-based murder mystery in which players question the ghost of a murdered clockmaker by placing keepsakes on a Día de los Muertos altar. The game runs Qwen2.5-1.5B…
OpenAI announced on October 26, 2026 that one of its internal AI systems produced a claimed refutation of the Exponential Time Hypothesis, publishing a manuscript and a Lean formalization describing a deterministic proce…
New York Governor Kathy Hochul signed a law on Wednesday that bars AI chatbots from offering a range of features to users under 18, including pretending to be human, giving flattering "people-pleasing" responses, and dis…
A benchmark of 5,276 graded runs across eight models on five small Pydantic AI apps found that code mode roughly halved cost and wall time on long tasks, cutting cost 58% and wall time 51% while raising success by 16.0 p…
Moonfold released v0.1.0-beta.1, an open-source MIT-licensed workspace that runs coding agent CLIs including Codex, Claude Code, Cursor, Grok Build, OpenCode and Antigravity locally with the user's own logins, and adds p…
Attackers are increasingly abusing legitimate AI services and agentic systems rather than malware, a tactic security experts call "living off the AI land," according to CSO. In September 2025, attackers published a count…
Requesty launched an AI gateway and LLM router that gives developers access to more than 600 models through one OpenAI-compatible API, with EU data residency in Frankfurt on AWS eu-central-1 via router.eu.requesty.ai. Re…
Three data platform engineers at a roughly 500-person company built an internal platform for agent-written HTML dashboards that reached 156 weekly users in the week of 5 October, against Looker's 157, according to David …
A developer outlined a pattern for classifying sales-call transcripts into CRM actions by embedding a small set of labeled examples and assigning each new transcript or chunk to the label with the highest cosine similari…
Microsoft created the public GitHub repository microsoft/aspire-templates on October 28, 2026, housing six Aspire starter templates and two repo-local agent skills that instruct an AI coding agent how to create and revie…
Artificial Analysis released Intelligence Index v5 on October 27, 2026, with Claude Opus 5.5 (max) leading at 58, one point ahead of GPT-6 Astra (max) at 57. The rebuilt benchmark averages nine evaluations across four eq…
GitHub made Copilot Fusion generally available on October 28, 2026, at its Universe conference, replacing per-task model selection with three effort levels — Standard, High and Max — that let Copilot route and orchestrat…
The MIT-licensed jakubkrehel/skills repository offers 14 agent skills that help coding agents review and build user interfaces, covering UI polish, typography, color, accessibility, layout and product writing, and had ab…
MiniMax-AI published mcode-stack, a public GitHub repository containing 15 agent skills and checklists that the MiniMax Code team uses to build MiniMax Code, with each item packaged as a SKILL.md an agent runs as an audi…
A Bakeoff benchmark of 5 coding agents across 5 codebases found that raising effort levels roughly doubled the amount of code changed but only 2 of 5 models solved more tasks, with Fable 5.1 performing worst at its high …