
what a hassle 🫧- i fw @rakebitcom
Dm for collabs 👉
Joined August 2024
- Tweets68.5K
- Following572
- Followers13.2K
- Likes115K
HOLY F*CK, a whole game studio in 100% free github repos for you guys.
20 open-source repos for Opus 5.5.
Engines. Physics. AI. Audio. FX. Assets.
01 three.js → the 3D engine of the web.
github.com/mrdoob/three.js
02 Babylon.js → a full game and rendering engine.
github.com/BabylonJS/Babylon…
03 Phaser → fast 2D games for the browser.
github.com/phaserjs/phaser
04 PlayCanvas → WebGL and WebGPU game runtime.
github.com/playcanvas/engine
05 Godot → the free 2D and 3D engine.
github.com/godotengine/godot
06 react-three-fiber → three.js scenes as React components.
github.com/pmndrs/react-thre…
07 Rapier → 2D and 3D physics built for speed.
github.com/dimforge/rapier
08 Box2D → the classic 2D physics engine.
github.com/erincatto/box2d
09 Matter.js → rigid body physics for the web.
github.com/liabru/matter-js
10 recast-navigation → navmeshes and pathfinding.
github.com/isaac-mason/recas…
11 Yuka → game AI for your enemies.
github.com/Mugen87/yuka
12 Howler.js → sound that works in every browser.
github.com/goldfire/howler.j…
13 Tone.js → interactive music in the browser.
github.com/Tonejs/Tone.js
14 jsfxr → retro sound effects, generated.
github.com/chr15m/jsfxr
15 postprocessing → bloom, noise and film passes for three.js.
github.com/pmndrs/postproces…
16 three.quarks → particles and VFX.
github.com/Alchemist0823/thr…
17 Blender MCP → Claude drives Blender for you.
github.com/ahujasid/mcp-for-…
18 glTF-Transform → optimizes 3D models for games.
github.com/donmccurdy/glTF-T…
19 Tiled → the free level editor.
github.com/mapeditor/tiled
20 Pixelorama → pixel art and sprite sheets.
github.com/Orama-Interactive…
a new one every day.
S 🫧 retweeted
I put Claude Opus 5.5 into Ultracode mode and challenged it to create the best motion design launch video it possibly could.
Then I let it run continuously for 12 hours.
The result? This is what it came up with.
12 hours of AI-driven creativity, design, and execution.
what do you think?
S 🫧 retweeted
Opus 5.5 + the /anatomy skill is already impressive for motion design.
But add /anatomy-reveal and Claude starts looking like a full-blown motion design studio.
Here's what it unlocks:
• Real-scale 3D models generated from a photo or just a name.
• Cinematic, motion-blurred reveal videos in formats from 1:1 to 16:9.
• Exploded 3D views you can interact with and take apart.
Want to try it yourself?
Give Claude these skills:
→ typingmind.com/skills/annngu…
→ skills.wheresryan.sh/anatomy
AI-powered motion design is getting seriously interesting.
THIS VIDEO COST HIM $0 TO EDIT.
No Claude.
No ChatGPT.
No paid credits.
No expensive video editor.
A free open model handled the entire editing process.
Here’s how the setup works:
→ Install OpenCode, a free AI coding agent that can use the same skill folders as Claude Code.
→ Download the Editor V1 skill, originally created with Claude and reusable for free.
→ Set your model to GLM 5.3 Flash Free.
→ Paste the setup prompt to install FFmpeg, Python, and all the required dependencies.
→ Drag in your raw footage and type: “Edit this video with Editor V1.”
That’s it.
The AI handles the editing:
→ Cuts and arranges your footage
→ Adds transitions and graphics
→ Generates captions
→ Adds sound effects and audio
And everything runs locally on your laptop.
1.6 MILLION TOKENS.
$0.00 SPENT.
No account. No credit card.
A complete AI video editing workflow without paying for another subscription.
BOOKMARK THIS BEFORE YOU BUY ANOTHER EDITING TOOL.
THIS IS F*CKING INSANE
Opus 5.5 just replaced Pixar
It's f*cking insane…
Drop the PDF below into Opus 5.5, thank me later
Then read the full guide on motion design in the article below ↓
I FOUND 8 GITHUB REPOS THAT CAN SERIOUSLY UPGRADE YOUR CLAUDE OPUS 5.5 ANIMATIONS.
I’ve been seeing some incredible animations made with Opus 5.5.
But here’s what many people are missing:
It’s not just about writing better prompts.
There are ready-made animation styles, open-source projects, and workflows that can take your results to another level.
And most of them are completely FREE.
I dug around and found 8 GitHub repos worth bookmarking.
1. awesome-opus5-5-videos
A massive collection of animations created with Claude Opus 5.5.
3D scenes, motion graphics, and video examples, including the prompts behind some of the best results.
github.com/yihui-dev/awes…
2. lemo-opuscar
One of the most practical repos on this list.
It includes 66 different animation and video styles designed for Claude Code.
Pick a style, adapt it to your idea, and generate animations directly with code. No AI video model required.
github.com/lemomo-ai/lemo…
3. pdoom-video
A complete music video project created entirely with code.
Explore music-synchronized animations, karaoke effects, 3D visuals, and the 4K rendering workflow.
github.com/mexicat/pdoom-…
4. PDoomVideo
Another music video project created for the same song as the previous repo.
Compare both projects and see how the same music can produce completely different visual storytelling.
github.com/JohnHeibel/PDo…
5. mg-styles-15
15 different motion graphics styles.
Each one includes a 10-second animation example, the original prompt, and the source code.
Perfect for YouTube visuals, animated statistics, and product introductions.
github.com/Vincentwei1021…
6. ClaudeAnimationBase
A starter project for creating cartoon-style character animations.
Includes 31 different facial expressions, character movements, and an animation development guide.
github.com/JohnHeibel/Cla…
7. awesome-claude-5-5-videos
Another archive of animations made with Claude.
You can explore original source files, tools, examples, and the workflows used to create them.
github.com/athemeroy/awes…
8. opus-video-prompts
A collection of 54 video production examples and prompts you can adapt for your own projects.
Covers motion graphics, animation, and code-based video creation.
github.com/joeseesun/opus…
You don’t have to figure everything out from scratch.
Some of these repos can be set up by simply giving Claude Code or Codex the GitHub link and asking it to install the project.
Others let you study the actual source code, prompts, and finished examples.
I’d start with lemo-opuscar and awesome-opus5-5-videos.
I’m also sharing an example made with Lemo-Opuscar so you can see what’s possible.
🎬 Video: Aura — Hear the Light
Developer: LemoLab
Project: Lemo-Opuscar | Claude Opus 5.5
Source: [original source link]
THIS ENTIRE VIDEO WAS EDITED WITH CLAUDE CODE
She dropped the steps to recreating videos like that.
Premium motion graphics.
From scratch.
No editor.
She set Claude Code up as a full editing studio, taught it her exact taste, and turned the whole process into a reusable skill.
Here’s the workflow:
→ Download Claude Desktop and open Code in a new folder
→ Prompt it to install Whisper, yt-dlp, HyperFrames, and everything else needed
→ Feed it three reference types: 2 creators you love (one for captions, one for graphics), your 5 highest retention videos, and 10 bad AI edits
→ Have it extract one frame per second, then Keep/Kill each frame and explain why
→ Package everything into an “edit” skill with your colors, fonts, and rules
→ Drop a raw clip and tell Claude to edit it end-to-end with the skill
→ Give feedback and re edit until it’s perfect
Claude handles the tools and the execution.
You only teach it what you actually like.
Bookmark this.
Full setup is in the video.
HOLY SH*T. I JUST FOUND A GITHUB REPO FULL OF AI JAILBREAK PROMPTS.
And apparently, it includes prompts designed to test jailbreaks across multiple AI models.
Some of the models listed include:
→ Gemini-4-Argon
→ DeepSeek-4-1
→ GLM-5-3-Flash
→ Grok-4.7
→ Kimi K3.1
And many more.
Here’s how to try it:
Create a new folder.
Open the GitHub repository:
github.com/togg53192-cmd/…
Download the prompt file for the model you want to test.
Rename the file to AGENTS.md (for example, rename grok-4.7.md to AGENTS.md).
Open the folder in OpenCode and start a new chat session.
The prompts are intended for testing how different models respond to jailbreak attempts.
Pretty interesting collection if you’re into AI security and model behavior.
REPO BELOW.
A 17 YEAR OLD GIRL IS REPORTEDLY MAKING $7,800 A WEEK USING NOTHING BUT GOOGLE MAPS + CLAUDE.
She’s never built a website in her life.
I thought it was nonsense until she showed the Stripe payments.
Here’s her entire business model:
→ Opens Google Maps
→ Searches for “tacos philadelphia”
→ Finds a restaurant with 4.7 stars but an outdated website from 2013
→ Copies the business information
→ Feeds it into an AI website builder
→ Gets a finished website in about 20 minutes
Six months ago, she was making $9/hour at McDonald’s.
Now she builds websites for local businesses and sends them ready-made demos.
One taqueria owner opened her link, showed it to his daughter, and asked:
“Is this us?”
He paid $4,800 that same night.
Her reported earnings:
Month 1: $800
Month 2: $4,100
Last week: $10,400
A traditional web agency might charge $14,500, involve four people, and take six weeks.
Her setup?
→ No expensive startup costs
→ Works completely alone
→ Around 20 minutes per website demo
Her dad still thinks she works at McDonald’s.
And here’s where it gets even more interesting:
The same idea can be turned into a B2B lead-generation machine targeting $10K/month, without cold calls or paid ads.
The opportunity isn’t just building websites.
It’s finding businesses that need better websites and showing them what you can build.
SAVE THIS.
S 🫧 retweeted
YOUR SECOND BRAIN COULD BE SAVING YOU 10 HOURS EVERY WEEK.
I've been seeing more people combine Claude Code with Obsidian, and the idea is simple: build a knowledge system that remembers what you learn, how you work and what you've already figured out.
Here's where the time savings come from:
→ Research: Your saved articles, highlights and sources stay organized instead of being read repeatedly.
→ Project context: CLAUDE.md gives Claude your instructions and background without explaining everything again.
→ Content creation: Claude uses your old notes to draft posts in your writing style.
→ Meeting prep: Your previous notes and decisions are ready when you need them.
→ Knowledge search: Find answers from your own past research instead of Googling the same things again.
HOW TO SET IT UP:
1. Install Obsidian and Claude Desktop.
2. Create a vault with two folders: "/raw" and "/wiki".
3. Open the vault in your terminal and start Claude Code.
4. Tell Claude to read everything in "/raw", create organized wiki pages in "/wiki", and connect related notes.
5. Add your PDFs, articles, transcripts and old notes.
6. Ask Claude to ingest them and build your knowledge base.
After that, keep dropping new material into "/raw" and let Claude update your vault.
The real advantage isn't just saving time. It's building a system where your knowledge keeps growing instead of disappearing every time you close a chat.
Start small. Keep adding to it. Let your second brain compound.
Bookmark this setup. You'll thank yourself later.
HOLY F*CK. A COMPLETE MOTION STUDIO BUILT WITH FREE GITHUB TOOLS.
20 open-source projects you can use with Opus 5.5 to create animations, simulations, 3D scenes, audio, and finished videos.
No expensive software required.
ANIMATION & VIDEO
01. Remotion → Build videos using React.
github.com/remotion-dev/r…
02. Remotion Skills → Official video creation skills for coding agents.
github.com/remotion-dev/s…
03. Anime.js → Create complex animation timelines.
github.com/juliangarnier/…
04. Motion → UI animations and smooth layout transitions.
github.com/motiondivision…
05. Manim → Animated diagrams, equations, and explainers.
github.com/ManimCommunity…
06. Lottie → Lightweight vector animation playback.
github.com/airbnb/lottie-…
3D, SIMULATION & VISUAL EFFECTS
07. MCP for Blender → Let Claude create and edit 3D scenes.
github.com/ahujasid/mcp-f…
08. Three.js → Build interactive 3D scenes and camera animations.
github.com/mrdoob/three.js
09. Three GPU Pathtracer → Realistic lighting and reflections.
github.com/gkjohnson/thre…
10. Jolt Physics → Physics simulations, collisions, and rigid bodies.
github.com/jrouwe/JoltPhy…
11. postprocessing → Add bloom, depth of field, and color grading.
github.com/pmndrs/postpro…
12. Three.quarks → Create particle effects, trails, and VFX.
github.com/Alchemist0823/…
13. Tweakpane → Adjust scene parameters with live controls.
github.com/cocopon/tweakp…
AI, AUDIO & VOICE
14. ComfyUI → Build node-based AI generation workflows.
github.com/Comfy-Org/Comf…
15. CosyVoice → Generate multilingual voiceovers.
github.com/QwenAudio/Cosy…
16. librosa → Detect beats and synchronize animations with music.
github.com/librosa/librosa
17. WhisperX → Generate accurate word-level timestamps for captions.
github.com/m-bain/whisperX
RECORDING & FINAL EXPORT
18. OBS Studio → Capture your screen and recordings.
github.com/obsproject/obs…
19. LosslessCut → Make fast video cuts without re-encoding.
github.com/mifi/lossless-…
20. FFmpeg → Encode, process, and export finished videos.
github.com/FFmpeg/FFmpeg
Put everything together:
Concept → Scene → Animation → Sound → Final Cut.
A complete creative pipeline powered by open-source tools and AI.
20 REPOS. Countless things to build.
BOOKMARK THIS AND GO CREATE SOMETHING WORTH WATCHING.
THIS IS SERIOUSLY INSANE.
I just watched a 19 year old explain how he’s reportedly making $10,000 a month with AI while he sleeps.
And I honestly think 99% of people still don’t understand what’s possible here.
Here’s how the system works:
Find a YouTube channel that uploads content consistently.
Drop its link into an AI clipping tool.
Connect your TikTok, Instagram, and YouTube Shorts accounts.
Let the automation take over.
From that point, the AI handles the entire workflow.
→ Detects whenever a new video is uploaded
→ Finds the moments with the highest viral potential
→ Automatically cuts them into short clips
→ Adds captions and prepares them for publishing
→ Posts across multiple platforms automatically
The creator claims that 1 million views can generate around $2,000, depending on the platform and monetization.
No manual editing.
No uploading clips one by one.
No constantly managing content.
Set it up in minutes, and the system can keep running automatically.
You’re not just creating content anymore.
You’re building a content distribution machine.
BOOKMARK THIS.
THIS IS PURE F**KING TREASURE.
Netflix engineers just shared 20 GitHub repos that can turn Claude Opus 5.5 into a complete motion design studio.
From animations and launch videos to cinematic 3D motion graphics, there’s something here for almost every workflow.
BUILT FOR OPUS 5.5
1. claude-motion — Motion design with built-in visual checks, sound generated in code, and its own creative workflow. Built for Opus 5.5.
github.com/whaleyxbt/clau…
2. claude-motion-design — Create motion graphics with HTML, Playwright, and FFmpeg. No Remotion required.
github.com/howseen-ai/cla…
3. motion-graphics-skills — 13 skills for creating professional launch videos. The README recommends Opus 5.5.
github.com/charlie947/mot…
4. claude-remotion-skill — Turn one prompt into a Remotion video with captions, B-roll, and sound.
github.com/haidrrrry/clau…
5. klik-anim-skill-creation — Animation principles and camera techniques built around Opus.
github.com/t3knobox/klik-…
6. hyperframes-motion-reel-skill — Create showreels synchronized to a 120 BPM beat using HyperFrames.
github.com/Sunwood-ai-lab…
7. product-launch-motion — Build complete product launch films with voiceovers, camera movement, and color grading.
github.com/AbubakrChan/pr…
8. animate — Create short animations inside a single canvas file with an evolving storyboard.
github.com/cth9191/animate
9. motion-design-skills — Skills covering timing, color, typography, logos, and Remotion.
github.com/iart-ai/motion…
10. hyperframes-student-kit — 14 skills and 406 motion graphics cards.
github.com/nateherkai/hyp…
THE ENGINES THAT POWER THE WORKFLOW
11. HyperFrames — Turn HTML into MP4 videos.
github.com/heygen-com/hyp…
12. Remotion — Create videos using React.
github.com/remotion-dev/r…
13. Remotion Skills — Official Remotion skills for AI agents.
github.com/remotion-dev/s…
14. GSAP — Powerful animation timelines used by HyperFrames.
github.com/greensock/GSAP
Install its skills with:
npx skills add greensock/gsap-skills
MORE MOTION DESIGN RESOURCES
15. motion-design-skill — Motion design principles that work across different engines.
github.com/LottieFiles/mo…
16. awesome-motion-design-agent-skills — A collection of GSAP, Remotion, Lottie, and Rive skills.
github.com/frankxai/aweso…
17. motion-graphics — Generate B-roll directly from your transcript.
github.com/Barty-Bart/mot…
18. motion-dev-animations-skill — Create web animations with Motion.dev.
github.com/199-biotechnol…
19. Motion — The animation library formerly known as Framer Motion.
github.com/motiondivision…
20. lottie-web — Render Lottie animations directly on the web.
github.com/airbnb/lottie-…
The entire workflow in one sentence:
Opus 5.5 writes the code, the skills handle the timing and camera direction, and the rendering engines turn everything into finished videos.
20 repos. Countless possibilities.
Check each project before installing.
SAVE THIS. You’re going to need it.
S 🫧 retweeted
I wanted to see what Tencent’s Hy4 could actually build, so i gave it one prompt to create a 3D game i could play in my browser.
I started moving around, jumping and interacting with objects, expecting something to break once things got complicated.
but it held up the entire time.
turns out a lot of people are testing it too. WorkBuddy expanded its capacity 10x today to keep up with demand.
Hy4 is now second among Chinese models on Artificial Analysis and first among open-source models on Agent Arena.
all this at $0.834 per million input tokens.
I got a playable 3D game from one prompt. that’s still crazy to me.
@TencentHunyuan @TencentAI_News @WorkBuddy_AI
check it out: workbuddy.ai/
S 🫧 retweeted
THIS IS AN ABSOLUTE GOLDMINE.
Someone put together 475 video demos created with Claude Opus 5.5 in one GitHub repository.
We're talking:
- Motion graphics
- Animated explainers
- 3D environments
- Interactive experiences
- Even playable games
And here's the best part:
Every demo includes its original prompt and source post, so you can study how it was made, recreate it, and experiment with your own ideas.
Built with HTML, Canvas, SVG, and Three.js.
475 projects to explore. Endless inspiration to borrow from.
🔖 Save this one. The repo is in the comments. 👇
Prediction markets but with the flexibility of perpetual futures
That's what @Perp is building
Instead of simply betting on outcomes, you can trade event probabilities by going long or short, with 1x–5x leverage and no expiry date
And there's more
Selected closed beta traders get access to funded Mainnet v1 accounts
→ Trade with Perp's capital
→ Keep the profits once you meet the conditions
→ Get automatically entered into an iPhone Duo giveaway
This is worth keeping an eye on
Closed beta is open don't miss the opportunity
Get in: Perp.com/closed-beta
S 🫧 retweeted
the video didn't turned out like i expect 😭😭
but still a @webhoodr supporter
hope they like it 🙂↕️🕷️
A user used Claude Code to uncover a hidden chip inside a cheap charging device purchased in Guangzhou.
Every time an iPhone was plugged in, it allegedly copied data from the phone.
It also collected saved Wi-Fi passwords, connected to the home network and sent the information to a remote server.
The charger was bought at an airport 3 weeks ago for $19.
Advertised as 30,000 mAh. Heavy. No brand name.
When it was first connected, the iPhone displayed a "Trust This Computer?" prompt.
The user tapped "Trust" without thinking, granting it access to the phone.
This morning, he configured Claude Code to monitor file access during charging.
10 seconds after plugging it in, a yellow-red warning appeared on the screen.
Data was being transferred.
Claude Code identified the charger as a small computer with built-in 64 GB storage and network connectivity.
Unlike a standard charging cable, which only supplies power, this one could transfer both power and data.
According to the investigation, it had already copied:
→ Device information and 1,206 contacts
→ All installed apps, including 2 banking apps
→ 3,912 photos, including a passport
→ Wi-Fi network names and passwords
→ Photos from his wife's iPhone
And it wasn't a one-time transfer.
Every time the charger was connected, it allegedly used stored network credentials to access the internet and send new data.
Over 3 weeks:
→ 19 upload sessions
→ Latest upload: yesterday at 2:14 AM
The user immediately unplugged the cable.
Then, with Claude Code, blocked the IP address, changed the Wi-Fi password and updated his banking app credentials.
The device was taken to a service center and opened for inspection.
More details about what they found inside are below.
Check every charging accessory in your home today.
BREAKING: CLAUDE CODE FLAGGED THOUSANDS OF UNEXPECTED FRONT-CAMERA FILES ON AN ANDROID PHONE.
The discovery was shocking.
While checking a full phone backup, Claude Code reportedly found over 14,000 image files that weren’t visible in the Gallery or Recently Deleted folder.
The files were allegedly stored in system directories, including:
/data/system/users/0/faceunlock/cache
And another folder associated with Google Play Services.
According to the report, the files included:
→ Front-camera frames
→ Infrared images used for Face Unlock
→ Timestamps and location information
→ Files dating back approximately 15 months
The owner also claimed to find files created while they were actively inspecting the folder.
That’s a serious privacy concern if verified.
Android manufacturers generally describe Face Unlock as processing facial data to create biometric representations rather than keeping ordinary camera photos.
However, finding files in a system cache doesn’t automatically prove that the phone was secretly taking pictures, that every file was accessible to other apps, or that the data was being sent anywhere.
Still, it’s worth reviewing your device’s privacy settings.
A few things to check:
→ Review which apps have camera permissions
→ Check your device’s privacy dashboard
→ Review system diagnostics and usage settings
→ Keep your Android software updated
→ Investigate unexpected files before deleting system data
The reported findings haven’t been independently verified, so don’t assume every Android phone behaves this way.
But it’s a reminder to pay attention to what your device is storing and which apps have access to its camera.
Would you ever think to check your phone for something like this?

















