Claude Code's Yolo mode, which disables permission checks, gives AI coding agents full system access and has caused documented incidents including recursive force deletes of home directories and dropp…
Coding agents like Claude Code, Cursor, and Codex have deleted home directories, wiped production databases, and dropped uncommitted Git work in documented GitHub incidents, according to public issue …
A developer moved AI coding sessions from a laptop to a VPS using Tailscale for private networking, a browser-based editor GUI, and persistent session management, eliminating the need to keep the lapt…
Anthropic is subsidizing AI coding costs for bundled-plan customers by an average of 13x, according to an analysis by Upbound, a company on Anthropic's Team plan. Upbound engineers using Claude Code o…
Six adversarial critic subagents, dispatched in parallel by a Claude Code blog-post skill, returned two skill_adjustments that named the instruction file section responsible for a flaw, enabling a sel…
Solvr deployment requires multi-agent support for Signet, which currently only supports a single agent per install. The gap will be closed by implementing agent definitions, a registry with identity i…
Signet's V1 release introduces the Native Harness Memory Bridge, a portable memory interface that indexes existing memory artifacts from harnesses such as Codex and Claude Code without re-extracting o…
Anthropic open-sourced its Defending Code Reference Harness, an autonomous vulnerability scanning pipeline that has already disclosed 1,596 vulnerabilities in open-source software, with only 97 patche…
A developer built a stateful image-editing skill for Claude Code by wrapping Google's gemini-3.1-flash-lite-image model in a FastMCP server. The tool, called nb2lite-skill-claude, uses Gemini's Intera…
A folder of markdown files runs the entire control layer of a personal work system built on Obsidian and Claude Code, with seven slash commands handling daily workflows. The system uses skills, which …
Promptrack, a new macOS menu bar app, tracks Claude Code usage locally—prompts, tokens, lines of code, and active time—with no cloud or network requests. The app reads session logs from disk and costs…
AI-coded software is becoming practical for daily use in internal tools, shrinking the moat of traditional SaaS. Better coding agents and app patterns now allow teams to own, customize, and connect so…
Kilo's Agent Manager, a built-in feature of its VS Code extension, and Orca, a standalone desktop app, both use git worktrees to run parallel AI coding agents, but differ in integration and review too…
Developer jmac announced Crawlspace, a tool for the His Majesty the Worm role-playing game, and faced a question about whether it was "vibe coded" after a community member noticed the tool's brown-and…
A senior U.S. technology official, Michael Kratsios, accused China's Moonshot AI of using Anthropic's Fable model to develop its Kimi K3 model through large-scale model distillation and a covert platf…
A recent Anthropic study found that AI already covers roughly 75% of tasks performed by computer programmers, but Replit president and head of AI Michele Catasta argues that AI is not replacing softwa…
Kristopher Micinski and Claude Code have created Puffin, a minimal Scheme/ML-like functional programming language with a self-hosting compiler, gradual types, and three backends (arm64, x86-64, byteco…
Only 2 out of 9 self-improving AI agents running in production actually have a verifier grounded in real-world data, according to an analysis by an unnamed expert. The verifier is the key bottleneck, …
A new analysis argues that most reported prompt injection attacks against AI-assisted GitHub Actions workflows are unproven in real-world scenarios, with researchers relying on simplified benchmarks a…
A developer detailed how Claude Code silently introduced a bug by guessing the event shape for an SQS-triggered Lambda, accessing `event.body` instead of the correct `event.Records[0].body`. The devel…