Codex Code Review can now use custom repository rules in AGENTS.md. Start with a check your reviewers keep repeating. Keep it concise and scoped so Codex can catch repository-specific issues even when the relevant context is siloed.

Verify, structure, tokenize, manage. MPC's framework starts before blockchain enters the picture, with the mine and its raw source data. Professional reviewers do the judging. MPC connects who verified what, when, and how, into one traceable record. See how it works.

# Claude Code Features and Practical Usage 🤖 Don't want logs and search results cluttering your main conversation? Subagents work in their own context and return only a summary. You can carve out read-only reviewers and research specialists. 📌 Title and Feature URL Title: Subagents URL: 📝 Overview A subagent is a specialized AI assistant for a specific kind of task. It runs in its own context window with its own system prompt, tool access, and permissions, working independently and returning only a summary to the main conversation. This isolates verbose output and lets you constrain what it can do via tool limits. 🔧 How It Works - Built-in subagents include Explore (a read-only, Haiku-powered codebase explorer), Plan (a research agent for plan mode), and general-purpose (for complex tasks needing both exploration and changes). - Custom subagents are defined as Markdown with YAML frontmatter; only name and description are required. - Key fields include tools (allowed tools), disallowedTools (denied tools), model (sonnet/opus/haiku/inherit), permissionMode, skills, and memory. - Claude decides delegation based on the description; phrases like "use proactively" encourage eager delegation. - Subagents cannot spawn other subagents, which prevents infinite nesting. 🛠 Practical Usage - The /agents command opens a tabbed UI to create, edit, and delete subagents via guided setup or Claude generation (recommended). - Storage location sets scope: .claude/agents/ for the current project, ~/.claude/agents/ for all projects. The former checks into version control for team sharing. - Invoke explicitly via natural language ("use the test-runner subagent to..."), an @-mention, or run the whole session as one with claude --agent code-reviewer. - For a read-only reviewer, restrict tools, e.g. tools: Read, Grep, Glob, Bash. - Set memory: project to give a subagent a persistent memory directory that accumulates insights across conversations. 🎯 Use Cases - Defining a code reviewer that never modifies code, or a read-only research specialist for long investigations. - Isolating high-output work like running tests or processing logs, returning only the failing tests as a summary. - Running parallel research on independent modules in separate subagents. - Routing tasks to fast, cheap Haiku to control cost. ⚠️ Caveats - Each subagent starts with a fresh context and does not inherit conversation history or already-read files; it may take time to ramp up. - Many subagents returning detailed results can themselves consume significant main-conversation context. - Editing subagent files directly on disk requires a session restart to load them (changes via /agents take effect immediately). - Use bypassPermissions carefully, as it skips permission checks. #ClaudeCode# #DevTools#

A few days ago, ECHO Asia was an idea. Today it has a website. Vietnamese materials. Japanese materials. Separate landing pages. Localized demos. Community reviewers. Checkout. Delivery. And a global audience watching it come alive in real time. That does not happen because a company “targets a market.” It happens because people show up. Vietnam showed up. Japan showed up. And we decided to meet that energy with real work. We don’t speak every language perfectly. We are not pretending we do. But ECHO was never only about translation. It was about context. How business feels different in Vietnam. How trust moves differently in Japan. How a conversation changes before the answer ever arrives. That is what we built. Not another prompt pack. A conversational operating layer that can adapt across language, culture, and business reality. The first ECHO Asia community versions are now live. Every early buyer will receive revised versions for free as native speakers and real operators help us keep improving. Built with respect. Built in public. Built with the communities that believed in us first. 🌊 🇻🇳 Gửi Việt Nam: Cảm ơn các bạn vì đã đón nhận ECHO với sự cởi mở, năng lượng và rất nhiều góp ý thật lòng. Chúng tôi biết mình không đến từ Việt Nam. Vì vậy, chúng tôi không muốn giả vờ rằng mình hiểu mọi thứ. Điều chúng tôi muốn làm là lắng nghe, học hỏi và xây dựng cùng những người thật sự hiểu cộng đồng của mình. Phiên bản Việt Nam đầu tiên này là bước khởi đầu. Nếu có điều gì cần chỉnh sửa, chúng tôi sẽ chỉnh sửa. Nếu có điều gì có thể tốt hơn, chúng tôi sẽ làm tốt hơn. Cảm ơn Việt Nam đã giúp chúng tôi đưa ECHO Asia thành hiện thực. 🌊 🇯🇵 日本のみなさんへ： ECHO Asiaを温かく見守り、応援してくださり、本当にありがとうございます。 私たちは日本語や日本の商習慣を完全に理解しているとは思っていません。 だからこそ、最初から敬意を持って、丁寧に、そして学びながら作りたいと思いました。 この日本語版は、完成された終点ではなく、コミュニティと一緒に育てていく最初の一歩です。 表現、文脈、文化的な違和感があれば、ぜひ教えてください。 より自然で、より役に立つものにしていきます。 日本の皆さん、本当にありがとうございます。 🌊 Vietnam. Japan. We’re just getting started.

We tested Claude, GPT-5.2, and Gemini as scientific paper reviewers on @KurateOrg. When scoring arXiv papers on impact (1–10), GPT clusters around 7–8 and Gemini is similar. Claude Opus 4.6 uses the full range, making it a far better discriminator. See the distributions 👇

new @poteto-mode config: hardest code, judgment or vague intent: claude-fable-5-1-thinking-high hardest code, precisely specified steps: cursor-grok-4.6-xhigh-fast code, everything else: cursor-grok-4.6-xhigh-fast trivial mechanical edits: cursor-grok-4.6-xhigh-fast judgment and prose: claude-fable-5-1-thinking-high how explorer: cursor-grok-4.6-xhigh-fast how explainer: claude-fable-5-1-thinking-high how critics: claude-fable-5-1-thinking-high, cursor-grok-4.6-xhigh-fast why investigators: cursor-grok-4.6-xhigh-fast why synthesizer: claude-fable-5-1-thinking-high reflect judgment: claude-fable-5-1-thinking-high reflect tooling: cursor-grok-4.6-xhigh-fast reflect divergent: cursor-grok-4.6-xhigh-fast reflect synthesizer: claude-fable-5-1-thinking-high arena runners: claude-fable-5-1-thinking-high, cursor-grok-4.6-xhigh-fast arena cross-judge pool: claude-fable-5-1-thinking-high, cursor-grok-4.6-xhigh-fast interrogate reviewers: claude-fable-5-1-thinking-high, cursor-grok-4.6-xhigh-fast @bot helped me iterate on this based on my transcripts

Melville Bissell patented a carpet sweeper in 1876. In 1889 his wife Anna took over and became America's first woman CEO, building it into the largest company of its kind in the world and setting its early standards for trademarks and patents. BISSELL is still family-owned 150 years later. Every product is built from parts, and every part needs an approved name before engineering can start. That approval ran on Excel forms, applied by hand against rules that lived in the heads of people who have been there 40 and 50 years. We built the system that runs it now. 7,000+ parts submitted, 81.4% approved in seconds, and reviewers spend their time on the ones that need a decision. The full story: