Because of the Amazing support you all gave for the first Round Wylder (my step daughter) made it into the second round! You can vote once a day and some days are 2x votes ! I would love love love if any of you support her going to work with some of the best animal rescues in the world to just cast at least one free vote if you can! You can vote here! Not Ai related so sorry for that ! Wylder | Junior Ranger

Quick one for anyone running Claude Code on knowledge work rather than code. There's a built-in command, /insights, that reads your last X sessions and produces a report on how you're actually using it. Mine just landed and it's worth sharing. What it gives you: - The areas you spend most time in (mine: client emails, quoting, workspace reorgs, daily inbox rhythm, strategic research) - Your interaction style, with one sentence that nails the pattern - Where Claude misfired and why (mine: filling gaps with plausible guesses instead of verifying, pulling from the wrong source file, output formatting that won't paste cleanly into Outlook) - Specific CLAUDE.md additions to stop the same friction recurring - Features to try (Skills, Hooks, MCP servers) tied to your actual usage, not generic advice - "On the horizon" workflows you could build given your patterns For me it surfaced that my three-way workspace split was working, but my email-drafting conventions weren't documented anywhere Claude could see. For someone running a coaching practice it might surface that client-note structure is fine but session-prep is repeatedly being rebuilt from scratch. For a researcher it might flag that source-citation conventions are missing. For an agency, that brief-to-deck handoffs keep needing the same corrections. It reads the seams in your system, whatever that system is. That's exactly the gap ICM is meant to close, and /insights makes it visible. Takes a minute to run. Just type /insights. Worth dropping a screenshot of your "At a Glance" section in the comments if you give it a go, would be interesting to see the patterns across the group. 🙏

Image and video generation lived in apps. Claude could think about them. It couldn't make them. That gap closed today. Pushing-creation is the cinematic prompt methodology installed as a Claude Code skill. It teaches Claude to brief like a director of photography, not whisper at a model. v0.3.0 adds a keychain bridge. Provider keys live in macOS Keychain. Claude calls a CLI that reads them at runtime and passes them straight to the API. The key value never enters Claude's context. The result: every Claude Code session, in any workspace, can author style packs and generate frames end to end. The shift The methodology and the executor are in the same loop. Before: author a brief in Claude. Switch apps. Paste the brief. Generate. Switch back. Now: author the brief in Claude. Generate from the same session. The three steps Brief. `/frames-brainstorm` runs a DP interview. Claude reads your reference images and writes style.md live. STYLE_ blocks fight defaults. NEG_ blocks fight failure modes. Image. `/frames-gen --kind image --shot-n 1`. Claude reads the shot from your storyboard, calls Gemini or GPT-image. The PNG lands in `renders/`. Video, `/frames-gen --kind video --shot-n 1`. Same shot, video provider. Kling video lands in `renders/`. One session. One workspace. No app-hopping. The keychain mechanic. The OS holds the key. Claude holds the brief. Add each key once: `security add-generic-password -U -s "pushing-creation:gemini" -a "$USER" -w "your-key"` Claude calls `bin/frames-gen`. Inside that process, the key gets read into a Python variable, passed to the HTTP client, and never logged or printed. Claude's transcript stays clean. The OS guarantee: encrypted at rest, scoped to your user, accessible only with your authorisation. The unlock Every Claude session inherits this. Storyboard. Product shot. Social post. Personal portfolio. Same Claude. Same methodology. Different keys, but the keys were already there. Two repos clone once. From then on, every workspace is a generation studio.

if you know what slash commands are this post isnt for you --- when i say "you're not lazy enough" i mean this lovingly, baby bird you're doing unpaid labor community service for anthropics repeating a prompt is proof you found something reusable so use it. slash commands are how you get good by being lazy once you get a taste of what slash commands can do you're gonna go full humpback chunk on them, i promise. WHATS A SLASH COMMAND? - a command you save in your docs/ folder that you can call at anytime. - anytime you notice yourself prompting the same thing to an agent, copy it to docs/ and name it something short-n-sweet - use it WHEN TO USE A SLASH COMMAND? YOU KEEP MAKING BUDDY WEAR THE SAME HAT You keep giving him an _identity_ like "you're a workspace auditor with..." then keep re-asking him to look for layer confusion, bloated docs, stale shit any rules u thought u added but didnt - all that stuff. Now it's /audit-context YOU KEEP ASKING LOOK BEFORE TOUCH if you keep saying "look around first before touching anything." "find the root docs, context files, major folders, active work, stable refs, and weird routing instructions" blah blah blah blah save ur fingers papa that's /map and move on. YOU KEEP SAYING DO NOT EDIT YET if your always sayin shit like "don't edit yet, just look, this is a read task only" you just need to send your little buddy on a /recon mission. u needed some space after the fifteenth shit iteration, we've all been there king. YOU KEEP ASKING FOR THE SAME PLAN if you keep asking "what should we change and why?" /propose-cleanup make it explain the smallest, safest fix, what files would change, what it prevents, and what it absolutely will not touch. this is when you need a clipboard instead of a chainsaw YOU KEEP REPEATING THE SAME PERMISSION RULES if you have to tell him over and over to only touch things you approve, just make it /apply-approved, outline the exact plan, what to touch and not touch, to stop if its ambiguous, report every file touch.

https://tinyurl.com/kidbiohw-1 My Kiddo was asking me to help studying for High School Biology. Easy and light work right. Well let's add context and nuance. 1. My kid spends time between two homes and trying the telephone or screen sharing version of tutoring is an added challenge. 2. Add to that I have to spend the time learning not only what the course material is but how it's being taught to kids (Thanks Florida School system). Not all of us have great teachers like Mrs. Van Clief. So I created a process to take all the lessons and make my kid a study guide. I didn't just throw this at Claude without a plan. I knew what outcome I wanted and did UAT with the customer (my kid). So I build her a Biology Study Buddy. What it is: A self- contained biology study app for students topic-based lessons with reading sections followed by multiple-choice quizzes. Here is my very complicated tech stack (sarcasm) Backend: Google Apps Script (.gs files) — Code.gs serves the web app via doGet(), Curriculum.gs holds all lesson content as a JS object. Frontend: Single-page HTML/CSS/JS — Index.html is the shell, App.html contains all styles, markup, and logic inline. Hosting: Deployed as a Google Apps Script Web App (free, no server needed) Data/State: Local Storage for student progress — no database, no sign-in Fonts: Google Fonts — Baloo 2 (headings), Fredoka (subheadings), Inter (body) Design: Dark theme (#0f1729 base), card-based topic grid, CSS custom properties, mobile-responsive Architecture: Extremely simple — 4 files total. The curriculum is a large structured JS array (topics → sections → quizzes with explanations). Apps Script's Html Service templates the HTML and serves it. No build step, no dependencies, no framework. Content: Each topic has an intro, multiple reading sections, and a 10-question quiz with answer explanations. Currently includes at least 3 topics (lab safety, the brain, nature of science) with the curriculum file being quite large (~38K tokens), so likely many more.