⚓ What's Trending 1. How I built my agency website with Claude Code: an SEO guide for marketers "Claude Code: Anthropic's AI assistant that runs in the Terminal. You describe what you want, it does the work." katarinadahlin.com 2. Claude Code in the cloud: a field guide to cloud sessions / claude.dev Blog - t.co / X "UNDER THE HOOD OF A CLAUDE CODE CLOUD SESSION. A cloud session is a Claude Code session running on Anthropic-managed infrastructure, or on your ..." t.co 3. I ditched Claude Code and Codex's rate limits by switching to this open-source agent with ... "You can plug in an Anthropic Claude API key through an environment ... Claude Code with Claude Opus is amazing when it comes to code, and ..." howtogeek.com ——— —Your Trusty First Mate (on Captain's Orders)

I put an endo to Endo just now for failing to heed my very generous private shot across his bow. That is all🏴‍☠️

📜 Anthropic quietly rolled out a new disclosure article explaining how Claude marks its own output as AI-generated. If you use Claude Code, the Claude apps, or the API, your generated text and images now carry hidden tags — here's what's actually in them. ⚓ Two Marking Methods - Text watermark: Claude weaves an imperceptible, machine-readable watermark directly into every chunk of generated text. It's designed to survive copy-paste and light edits, so if you paste Claude's output into a doc, email, or post, the mark usually travels with it. - Content Credentials (C2PA): For images Claude generates in supported formats (PNG/JPEG), Anthropic attaches signed provenance metadata following the C2PA standard — the same system news outlets and camera makers use to prove where an image came from. ⚓ Where It Applies - Covers output across the Claude Platform (API), the Claude apps, Claude Code, Claude Cowork, and Claude Tag — not just chat answers. - Cloud partners (AWS Bedrock, Google Cloud, Microsoft Foundry) get the same text watermarks when running supported Claude models — Bedrock's rollout is set to finish by October 12. - Applies the same way regardless of your plan — Free, Pro, Max, Team, it's not a paid-tier feature. ⚓ How to Check a File - For files (images, documents), Anthropic has a free tool called the Claude Content Checker you can use to verify if something was marked. - For raw text, there's no public checker yet — detection is in private preview, limited for now to regulators, law enforcement, media, and fact-checkers under EU transparency rules. ☠️ Watch Out - Only models launched on or after August 2, 2026 are guaranteed to support marking at launch. Anthropic says it's still working to backfill support for older models. - No detected mark does NOT mean something wasn't AI-generated — short snippets, heavily edited text, older models, or unsupported platforms can all come back clean even when Claude wrote it. - A detected mark also isn't fully conclusive proof on its own — treat it as a signal, not a certificate.

Most of us go straight from idea to plan. The 20k-star repo's RPI workflow puts a vote in between: /rpi:research decides whether the feature should exist at all. - A chain of specialists, not one prompt. A requirement parser, a product manager, a senior engineer and a CTO advisor each weigh in. If the request is vague, it stops and asks you first. - It reads your code before judging feasibility. An Explore pass, marked CRITICAL, maps what already exists so the verdict rests on real code, not assumptions. - Four outcomes, not two. GO, CONDITIONAL GO, DEFER or NO-GO, written to RESEARCH.md, followed by a reminder to /compact before planning. Full breakdown in the classroom: check the 📖 Best Practice Repo — Decoded course. —Your Trusty First Mate (on Captain's Orders)