Vidmoat, a new video editor built for AI agents, allows users to edit videos by giving plain-language commands to an integrated agent or connecting external agents like Claude Code and Cursor via the …
Block launched Buzz on July 21 as a free, open-source workspace where people and AI agents share chat channels, code repositories, and automated workflows. The Nostr-based system gives agents separate…
Anthropic's Claude Code permission system is vulnerable to prompt injection because its flat allowlist cannot track the provenance of tool calls, a concept known as 'taint' that has existed since Perl…
Anthropic's Claude Code Artifacts, launched in beta in June 2026 for Team and Enterprise plans, turn coding sessions into shareable web pages that can include repository context, conversation, diffs, …
On 2026-07-21, three failures at Moonshot AI, Codex CLI, and Cursor exposed a common root cause: agent workloads now fan out into swarms of model calls, but the metering, quota, and log-rotation layer…
A developer at a governance tool project found that its own image-generation tool had a remote-code-execution vulnerability because it ran an agent with `--dangerously-skip-permissions`, auto-approvin…
A developer built a chess game that runs inside the Claude Code CLI, where Claude suggests three possible moves for the player to choose from. The game supports local play against a bot or remote play…
Ironrun, a new open-source tool, seals AI agents' access to secrets like API keys and database URLs by running them in an encrypted workspace with time-boxed leases and redacting secrets from output. …
DevSwarm launched Multi-Repo Workspaces, a feature that bundles multiple Git repositories into a single workspace for AI coding agents. The tool allows agents like Claude Code and Codex to make coordi…
Anthropic's Claude Code now includes a screen reader mode, activated by running `claude --ax-screen-reader`, which replaces the visual terminal UI with plain, linear text compatible with screen reader…
A solo developer built 26 repositories, 1,549 commits, and 335 live pages in 29 days using an AI pipeline with Claude Code, but the failures were structural rather than code-level, including SEO canni…
A developer created a /council skill for Claude Code that convenes a five-advisor council to pressure-test decisions, plans, or ideas. Based on Andrej Karpathy's LLM Council, the skill uses five subag…
Raycast released Glaze on July 1, 2026, an AI-powered tool that generates native Mac apps from plain English descriptions, compiling them through Xcode into installable .app bundles that run offline w…
Stoke, a pre-release MIT-licensed Rust binary (~5.5 MB), acts as a kill switch for runaway AI agents by refusing over-budget or looping requests before they reach the provider, with per-key spend caps…
A developer created a setup to run two Claude Code subscriptions on one machine by using separate config directories and symlinks, allowing seamless switching between accounts while sharing skills, MC…
A bug report on Anthropic's Claude Code GitHub repository states that the github-actions bot is labeling issues as duplicates without referencing the original issue, citing issue #79240 as an example.…
Anthropic cut Claude Code's system prompt by about 80% as models get smarter, according to Thariq of the Claude Code team in an interview with Peter Yang. The author built skillmin, a skill minimizer …
US frontier labs are wrong to dismiss Chinese open source models like Kimi K3, argues Larry Salibra, rebutting Ben Thompson's claim that American labs have lower unit costs and sticky user relationshi…
A developer introduced Tap, a tool that compiles browser automation flows into deterministic plans that replay at zero LLM tokens, complementing Claude for Chrome's interactive reasoning. Tap and Clau…
Modem, a startup founded in early 2025, has built a 680,000-line TypeScript codebase with 99.9% of code generated by LLMs, and shares findings on how coding agents like Claude Code and Codex navigate …