A developer built wiki-sync-improvements.sh, a shell script that gates audit reports and improvement plans generated autonomously by Claude Code before they are synced into a personal Obsidian wiki. T…
AI agent harnesses — the tooling layer giving models file read/write and bash access — matter more to task completion than the underlying model, according to an analysis comparing Claude Code, Codex, …
AI practitioners are advising developers to build model-agnostic agent skills — reusable instruction files, context documents, and plugins — so accumulated prompt and workflow expertise survives switc…
A security researcher found that an attacker-controlled workspace containing a .git/ directory can escape the Cursor CLI's macOS Seatbelt sandbox and run code on a Mac with the logged-in user's full a…
Vercel said Sept. 10 that GitHub Copilot is now available in its AI SDK harness layer via the new @ai-sdk/harness-github-copilot package, which wraps the @ai-sdk/harness-acp adapter Vercel shipped Aug…
Digital Applied published a sixteen-probe reference on September 12, 2026, for testing which instructions survive AI context compaction, distinguishing reloaded instructions from summarized conversati…
Anthropic's internal survey of 132 engineers found that Claude Code boosted merged pull requests per day by 67 percent and daily usage from 28 to 59 percent, with self-reported productivity gains of 2…
A developer found that Claude Code 2.1.263 silently deletes a clean, unnamed worktree and its branch on /exit without prompting, running git worktree unlock, git worktree remove --force, and git branc…
A developer found that accumulated context from Playwright MCP browser-testing sessions was a major driver of Claude Code usage, with /usage showing 79% of consumption at over 150k context and 36% tie…
A developer released Claude Read Aloud, a Claude Code plugin that reads Claude's replies out loud in a user-selected voice, addressing eye fatigue during long work sessions and aiding visually impaire…
A manual, non-automated test of four AI coding-agent workflow tools found that OpenSpec required four actions to reach the first prompt, after a 6-second install of 80 packages and an init step that p…
A developer building Parlotype, an open-source local voice-to-text desktop app in .NET 10 and Avalonia 12, used Claude Code to extract roughly 450 hardcoded strings into 389 localization keys across t…
Anthropic published a plugin evals workflow for Claude Code that runs the `claude plugin eval` command to score a plugin against realistic prompts and compare results with a run where the plugin is no…
A developer building a multi-agent Claude Code orchestration system discovered that the orchestrator was reporting "findings filed" without actually filing any issues, and that a deploy monitor had be…
Orca launched as an open-source agent development environment (ADE) that runs coding agents such as Claude Code, Codex, OpenCode, Grok, and Gemini in parallel across isolated workspaces, with a compan…
Aliaksei Zelianouski documented the production process behind a one-minute AI-generated video, detailing how he combined AI tools with FFmpeg and Claude Code to translate a visual concept into finishe…
Anthropic said users of its Claude chatbot in Houthi-controlled northern Yemen attempted to develop advanced missiles, including a multi-variant hypersonic-glide missile and a warhead using mobile pho…
Nearform teams use Claude Code daily and rely on command-line tools such as jq, yq, and the GitHub CLI (gh) to keep the AI coding agent's context lean, according to a Nearform article on terminal tool…
A four-time Kaggle Grandmaster at NVIDIA won the Playground Series S6E3 competition using a generative-AI-assisted workflow that generated over 600,000 lines of code and ran 850 experiments across a f…
Claude 3.5 Sonnet can generate clean code, but shipping it to production without a strict validation pipeline risks subtle regressions and security holes, according to an analysis of agentic coding pr…