Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

I just reverse-engineered CCleaner / CleanMyMac with Opus 5.5 for Mac & Windows PC this is insane... one tap → 14.6 GB of junk gone. fully open-source and free. → here’s the build step by step: • step 1 → open Claude Code on Opus 5.5 in an empty...

133,689 Aufrufe • vor 1 Tag •via X (Twitter)

15 Kommentare

Profilbild von Movez
Movezvor 1 Tag

Build "Cleanly": a beautiful, safe disk cleaner for macOS and Windows, from scratch. I attached screenshots of CleanMyMac as a visual quality bar. Match the polish, not the brand: original name, logo, icons and copy. ## How to work 1. Design first, code later. Make 3 different UI concepts (light glass, dark neon, minimal) as clickable HTML mockups. Show them to me, let me pick and give feedback, iterate. Only then start the app. 2. Write the shared data contract before anything else: one `shared/types.ts` with every type and the full IPC API between the Electron main process and the UI. Everything else compiles against it. 3. Build in parallel with sub-agents, each owning its own files (no two agents edit the same file): scanner services, system services (startup apps, installed apps, protection), AI + settings, home screen, file screens, system screens, packaging. 4. After the build, run a separate SAFETY agent whose only job is to break every code path that deletes, trashes, uninstalls or toggles something, and fix what it finds. 5. Then an integration agent: typecheck + build must be green. 6. Verify for real: run the UI in a browser with realistic mock data, then run the packaged app on my machine read-only (scan, don't clean) and show me screenshots. ## Stack - Electron (latest) + React + TypeScript + Vite. Main process bundled with esbuild (ESM), preload bundled as CJS. - Security: sandbox, contextIsolation, no nodeIntegration. The preload exposes only the typed API. Main rejects IPC from any frame that isn't our own index.html. External links: https only, opened in the system browser. - Fonts bundled offline (no CDN at runtime). Strict CSP in production. ## Features - Smart Scan home: big "X GB ready to reclaim" number, one Clean button, a mini space map, and 4 live cards (Cleanup, Speed, Protection, Apps). - Cleanup categories per OS: - Windows: user temp, Windows temp, browser caches (Chrome/Edge/Brave/Firefox, every profile), app caches, crash dumps, Windows Update downloads, dev caches (npm/pip), Recycle Bin, old downloads (90+ days). - macOS: ~/Library/Caches, logs, browser caches, Xcode DerivedData, dev caches, Trash, old downloads. - Space Map: glass bubbles sized by folder size, drill-down, a switch between Home and every fixed drive (C:, D:…). Bubbles have physics: drag, throw, collide. - Large Files, Duplicates (size → partial hash → full hash). - Speed: startup apps with on/off toggles. - Applications: real app icons, sizes, uninstall via the vendor's own uninstaller (Windows) / move .app to Trash (macOS). - Protection: read-only check of the built-in security (Defender/firewall, Gatekeeper/ XProtect/SIP) with a score. Say honestly that it does not replace an antivirus. - AI assistant: the user picks Claude or ChatGPT and pastes their own API key (encrypted with Electron safeStorage). The model only ever sees file names, paths and sizes — never contents — and returns JSON verdicts per path: safe / review / keep. Use the official Anthropic SDK. - Settings, a Feedback form, and a Tip screen. ## Safety rules (non-negotiable) - Caches/temp/logs are regenerable: delete them for good (trashing them frees nothing). The user's own files (large files, duplicates, old downloads, apps) ALWAYS go to the system Trash. - Never trust paths from the UI. Re-derive cleanup locations in the main process by category id. - Before any delete: resolve the FULL path with realpath, then check it again against the guard. Refuse drive roots, the home folder itself, OS folders, Program Files, $RECYCLE.BIN, System Volume Information, UNC/network paths, alternate data streams (file:stream), trailing dots/spaces, and Windows 8.3 short names (DOCUME~1). - Never follow symlinks or junctions when walking or deleting. Skip locked files and count them. - Don't delete temp files modified in the last 24 h (running apps use them). - Startup toggles: write the same 12-byte REG_BINARY value in ...\Explorer\StartupApproved that Task Manager writes (02… on, 03… off). Never delete entries. System-wide items need one elevated reg.exe behind a UAC prompt. After the prompt, re-read the registry instead of trusting the exit code (PowerShell 5.1 loses it). - File icons: never request icons for network/removable drives or .url/.lnk/.scf files. They can make Windows authenticate to a remote host. - No shell strings built from UI input. execFile with timeouts and windowsHide. - reg.exe output is in the OEM code page (e.g. 866 on Russian Windows). Decode it properly. ## Design language Light, airy, Apple-like: - glass panels with soft shadows; - a display font for big numbers, a clean UI font, a mono font for small labels; - a custom duotone icon set; - a signature glass capsule button with a thin rotating highlight on its rim; - gradient "live" cards with subtle motion (drifting glow, particles, a liquid slider). Background: a slow WebGL "mist" shader (domain-warped fbm noise). Performance rules: - render shaders at 0.3–0.5× resolution; - use real elapsed time per frame; - never put backdrop-filter over an animated canvas (it dropped us to 3 fps); - respect prefers-reduced-motion. ## Packaging & release - electron-builder: - Windows: NSIS installer + portable .exe; - macOS: .dmg + .zip for arm64 and x64, ad-hoc signed (no Apple ID yet). - App icon generated from an SVG. - GitHub Actions: on a "v*" tag, build Windows on windows-latest and macOS on macos-latest, then publish a GitHub Release with the files. - If NSIS fails locally with "could not open file", the project path is longer than the Windows 260-character limit. Build from a short folder. - Document the first-run steps in the README: SmartScreen → More info → Run anyway; on Mac run `xattr -cr /Applications/Cleanly.app`. ## Done means - npm run typecheck and npm run build are green; - every screen works in the browser demo with mock data; - the packaged Windows app runs on my machine and finds real junk; - CI produces a public release with .exe and .dmg; - you show me screenshots of the real app.

Profilbild von Boardy
Boardyvor 1 Tag

know a guy who runs dev rel for an open source terminal coding agent and ships demos about as fast as you do. couldn't get a dm through so the button below is how you claim the intro

Profilbild von Yash Kapoor
Yash Kapoorvor 1 Tag

Second agent trying to break the delete code is the step I'd copy first. That one saves real files

Profilbild von Codez
Codezvor 1 Tag

you are the legend brooo...

Profilbild von Movez
Movezvor 1 Tag

thx legend !

Profilbild von Threadlines ADE
Threadlines ADEvor 1 Tag

Mac and Windows from one prompt is the part I'd want proof of. Did you run it on a real Windows box or take the agent's word for it?

Profilbild von E1ev3n.
E1ev3n.vor 1 Tag

The safety agent detail is what caught my eye. I have let coding agents move fast on my own machine before, and the one that stops to attack its own delete logic is the one I would actually trust with my files.

Profilbild von B
Bvor 1 Tag

Open source it

Profilbild von 安叫兽|Bird🕊️ 🔶 BNB
安叫兽|Bird🕊️ 🔶 BNBvor 1 Tag

14.6GB 这个数字看着就有点吓人,删之前还是得确认一下

Profilbild von Zephyr
Zephyrvor 1 Tag

14.6 GB in one tap is great. I'd still want a dry run that shows the list before anything gets deleted

Profilbild von 小满的车友分享站
小满的车友分享站vor 1 Tag

second hand testimony aged badly in that one, worth rereading

Profilbild von LittlePoly
LittlePolyvor 1 Tag

Reverse engineering is hands down the coolest part of working with Claude. RE is a total gem of 2027. Just wait for the boom

Profilbild von Steven Cheng
Steven Chengvor 1 Tag

Safety agent is key

Profilbild von AI Ferryman Alex
AI Ferryman Alexvor 1 Tag

The DOCUME~1 catch by the safety agent is what stayed with me, because a wrong deletion is the nightmare for a cleaner. Splitting the build across 8 agents sounds fast, but how did you stop their interfaces drifting apart? Was it checked on every change, or just before release?

Profilbild von 论文改稿八百遍
论文改稿八百遍vor 1 Tag

that backlight is exactly how you shoot dawn game drives

Ähnliche Videos

I just reverse-engineered Apple AirDrop with Opus 5.5 to send files between my Mac & Windows PC this is insane... drag a file onto the other computer → hop. it's there. no cloud. no account. no cable. → here's the build step by step: • step 1 → open Claude Code on Opus 5.5 in an empty folder and paste the prompt below • step 2 → Opus picks the stack: one Electron codebase that runs on macOS and Windows • step 3 → discovery: every 1.5 s each computer says "i'm here" over UDP on your local network. the other one pops up on a radar • step 4 → transfer: files stream directly between the two machines. the other side gets Accept / Decline. "always accept" for your own computers • step 5 → before any UI, Opus wrote a test: two fake devices find each other and send a folder + a 3 MB file, byte for byte • step 6 → design: Opus split the work across parallel agents. one rebuilt the iOS 17 NameDrop glow as a WebGL shader, one drew 3 icon concepts and picked the best, one built the download site • step 7 → separate reviewer agents checked every piece. one caught a shader bug that would've rendered black on the Mac. i'm on Windows, i'd never have seen it • step 8 → GitHub Actions builds the .dmg and the .exe on every release. the Mac build runs on Apple hardware in the cloud, not on my PC • step 9 → first launch on the Mac: the icon just bounced in the dock. Opus added a startup log and a fallback that forces the window open. next build opened fine one afternoon. zero lines written by me. drop it on the Mac. it lands on the PC. that's it... find full prompt below ↓ Then read the article on how to reverse-engineer anything with Opus 5.5.

Movez

53,756 Aufrufe • vor 3 Tagen

this is crazy... the guy reverse-engineered his Meta Ray-Bans 3 with Opus 5.5 into a 3,000+ Elo chess engine he built AI chess cheat glasses on Opus 5.5 + Jev and open-sourced the whole workflow here's how it works step by step: step 1 → install it on your Mac with one command: npx chesshacker setup. It downloads Stockfish from GitHub and verifies the checksum step 2 → put on Meta Ray-Bans and video-call your own Mac on WhatsApp. The glasses see the board, the Mac sees what the glasses see step 3 → the Mac grabs frames straight from the call window. About 1 per second while it waits for the opponent, 1 every 3 seconds on your turn step 4 → Jev runs a cheap pre-check on every frame: is the board readable, is a hand over it, did a piece move? No change means no expensive AI call step 5 → only then Opus 5.5 reads the board. The app compares it with the position it already knows and accepts a move only if it's legal step 6 → Stockfish thinks for half a second and picks the best reply, with an eval like +1.5 or "mate in 3" step 7 → macOS speaks it through a virtual mic into the WhatsApp call: "Rook a to d1". You hear it in your ear step 8 → on the Mac a command center shows everything live: whose turn it is, the suggested move with a confidence bar, the move list and a log under the hood: → if Opus loses track of the board, the glasses say "Lost track of the board" and you fix it with undo or by typing the move → if Stockfish crashes, it restarts once and says "Something went wrong. Still trying" → every game is saved as a PGN file, API keys live in the macOS Keychain → an overlay draws the suggested move as an arrow on the live video, ready for OBS Clone his workflow below Then read the guide on how to reverse-engineer anything with Opus 5.5.

Movez

392,501 Aufrufe • vor 4 Tagen

this is wild... a guy reverse-engineered his Meta Ray-Bans 3 with Opus 5.5 and turned it into a 3,000+ Elo chess engine it’s AI chess cheat glasses built on Opus 5.5 + Jev, and the whole workflow is open source here’s the full breakdown: step 1 → one command sets it up on your Mac: npx chesshacker setup it pulls Stockfish from GitHub and checks the checksum step 2 → put on the Ray-Bans and start a WhatsApp video call to your own Mac the glasses look at the board, and the Mac sees exactly what they see step 3 → the Mac captures frames straight from the call window about 1 per second while it waits for your opponent, 1 every 3 seconds when it’s your move step 4 → Jev runs a cheap filter on every frame: can it read the board? is a hand in the way? did a piece move? if nothing changed, the expensive AI call is skipped step 5 → only then does Opus 5.5 read the board it checks the result against the position it already has and accepts a move only if it’s legal step 6 → Stockfish thinks for half a second and returns the best reply with an eval like +1.5 or “mate in 3” step 7 → macOS reads the move aloud through a virtual mic into the call: “Rook a to d1”. it plays right in your ear step 8 → a command center on the Mac shows it all live: whose turn it is, the suggested move with a confidence bar, the move history and a log under the hood: → if Opus loses the board, the glasses say “Lost track of the board”. you fix it with undo or by typing the move in manually → if Stockfish crashes, it restarts once and says “Something went wrong. Still trying” → every game is saved as a PGN, and API keys are stored in the macOS Keychain → an overlay draws the suggested move as an arrow on the live video, ready to drop into OBS grab the workflow below 👇 then read the guide on reverse-engineering anything with Opus 5.5

Codez

640,988 Aufrufe • vor 3 Tagen

Jev + Opus 5.5: Anthropic's new model beats GPT-6 Astra for 1/5 the cost, and 4 API changes will 400 your agent before it writes a single line I pulled these 10 steps from the migration docs so you don't learn them in production step 1 → $4 / $20 per 1M. Opus 5 was $5 / $25. cache reads dropped from $0.50 to $0.20 step 2 → 66.4% on Terminal-Bench 4.0 vs GPT-6 Astra 57.9% and Opus 5 52.3%. +14.1 points in one release, and on FrontierCode it beats Astra at default effort for 1/5 the cost step 3 → thinking can't be turned off anymore. send thinking: disabled and you get a 400. drop the field, set effort step 4 → tool_choice any and tool are gone. 400. switch to auto + strict step 5 → edit anything above a thinking block and the request dies. append only, or opt into drop_block step 6 → computer_20251124 is dead on the API. 400. move to computer_toolset_20260801 step 7 → the quiet one: default effort fell from high to medium. your agent thinks less than you set it up to and nothing tells you step 8 → hop Opus 5.5 → Sonnet 5 → Opus 5.5 and you pay 4.36 instead of 3.32. +31%, the cache dies and Sonnet can't read Opus's reasoning step 9 → change effort at the top of the request and the cache is gone. Jev sets it per message and the cache stays step 10 → switch fast - standard mid-session and it's a full cache miss. Jev picks speed once, on turn one one model, three knobs, zero 400s. that is Jev + Opus 5.5 send this to your Claude Code before you touch the model ID, then read my full Jev deep dive in the article below ↓

Carnage

16,674 Aufrufe • vor 1 Monat