Machine

Before you start: read this whole install prompt, tell me in a few lines what this skill does and what installing it involves, and wait for my go-ahead before you do anything else. Prerequisites: none for the install itself: no account, no API key, no paid service, nothing to install first. Using the skill afterwards needs Node 18 or newer, Playwright with Chromium, and ffmpeg (the filming and screenshot scripts); the pages the skill generates load three.js and their fonts from CDNs when opened, so viewing one needs an internet connection. All of that is use time, never this install. Install "machine" for me. It turns one line naming a machine ("a turbofan jet engine", "a mechanical watch") into a photoreal, physics-based, interactive 3D simulation of that machine in a single HTML file, with a small solver computing the real numbers on screen, plus a 20 to 30 second film of it. Repository: https://github.com/AirsupHQ/machine-skill (the skill is the whole repository at the pinned commit: SKILL.md at the root, with kit/, references/, scripts/, examples/ and film/) Pinned version: install the code at commit e0bd066ac5beaf48512935de2fc0f0e55636853a exactly. Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Clone https://github.com/AirsupHQ/machine-skill into a temporary folder, check out commit e0bd066ac5beaf48512935de2fc0f0e55636853a, and copy the whole repository folder (everything except the .git folder) from that checkout. Then compare every file with the fingerprints recorded for this listing on 2026-10-19: [FINGERPRINTS] - .github/REPO.md: SHA-256 e4825672881ce0dfc7e2a43cd78a6a1e92dfc68e88985cab87335cfbab0cb680 - .github/assets/compressor.jpg: SHA-256 3310150b8ea1f2b54bb4dce986bc14fa2627c4731bcbda6282ae92d7a8e1f68a - .github/assets/hero.webp: SHA-256 de0e3244b86c47dd8a54516c24d822fcacc762db099a5e9a3b6492ec07ba1f8d - .github/assets/refinery-pump.jpg: SHA-256 61e8dd1538f9433130fc70c0611f2de6709e8fc282f504a39ec5267fd98cdf19 - .github/assets/rocket-engine.jpg: SHA-256 cf1bf6df94af84ded6503dd8b51b6b06a987e9ccf9a7a3f66c9e7af724e1c3e8 - .github/assets/social.jpg: SHA-256 0d6e904b8ef3625acd8f4fbb2388742dab6a06bafdcd092bd3fd5f56b06d0bc1 - .github/assets/turbofan.jpg: SHA-256 a5409da9037c408aeee6acfca0544e88447cd026c5426902ee4244ccab6d056e - .gitignore: SHA-256 76d7986eaf13f636a42c86179bbfcce28b9d25eb0c841ebd734be694837f5e8a - CHANGELOG.md: SHA-256 18b9f7637b34495c5a63d610f2245cdb7f0a56a810f42b1b9296c423e54255d5 - CONTRIBUTING.md: SHA-256 f6ae649741e70b0802e85fd7e8d2cd8b502fb968c7d1ce76a12832e5a5d9c72a - LAUNCH.md: SHA-256 6cde4454c26605507119b13a2983f4f20dfae640249744264c18ad931629ab4a - LICENSE: SHA-256 7c20e0daf6c82af38eb9ce32f4f78ff61fc2172c1914d0fbec80c26c44fd4143 - README.md: SHA-256 3e4bf4c1de66bed0f0ac4277522d7f172f826a78937e41c7e8840298cb86961d - SKILL.md: SHA-256 1f71d2c90c9b1c0cca4b7af526b5a9721cb039638e872fb959173a5ec533611c - V2-IDEAS.md: SHA-256 dcb471193546df5b07ef062b2cd38bcace551528b8483a8ca652dadcb7c31628 - evals/evals.json: SHA-256 3f270da154ae6d9491eae35f7024e6d6812888f7f096b91a5ac279dde0c772af - examples/compressor.html: SHA-256 297313fe26a050a0308335c00120bdb9d46b66ff10aaa07e0d3eddf17feb51ef - examples/posters/compressor.jpg: SHA-256 3310150b8ea1f2b54bb4dce986bc14fa2627c4731bcbda6282ae92d7a8e1f68a - examples/posters/refinery-pump.jpg: SHA-256 61e8dd1538f9433130fc70c0611f2de6709e8fc282f504a39ec5267fd98cdf19 - examples/posters/rocket-engine.jpg: SHA-256 cf1bf6df94af84ded6503dd8b51b6b06a987e9ccf9a7a3f66c9e7af724e1c3e8 - examples/posters/turbofan.jpg: SHA-256 a5409da9037c408aeee6acfca0544e88447cd026c5426902ee4244ccab6d056e - examples/refinery-pump.html: SHA-256 8f58383a4bb2cb776a4dd0a83e2adcd98df40b16223a9cd0b173721229a71264 - examples/rocket-engine.html: SHA-256 60e90c24cf88ada8bc2fdc09a14702654b911e7f54151223b4edc8101667f154 - examples/turbofan.html: SHA-256 1129438d13e4d8ab080d485c5a4077865815f751c9cd5216be5817c6e96d2bfa - film/README.md: SHA-256 b5bc9e0fbfdc003e6c68746d04fc01614ba664417aa13e8e15104d77903c259f - film/launch.html: SHA-256 74d602c2f28a1b8485d4545a7b6fdf5de5f29256163839cae3897f3a3c382a48 - film/reply.html: SHA-256 1b59046e227b88e48d9fe1bb7317ae1dd29c874db312f9487f7f75059f2fd404 - film/stills/compressor_0.0.jpg: SHA-256 28e83cbd52e36bd73b5bffc2c543ef89c8b7a6cdfc42ca677a669c0a109c8fa5 - film/stills/refinery-pump_0.0.jpg: SHA-256 31b67440508551b4cbbde2d84e34df482db944190b3c05f2087bc27e968be961 - film/stills/rocket-engine_1.5.jpg: SHA-256 a2d14a29cebc0f6351788c11d838b6245d3ae18c956c510f75ea57a1d5af2420 - film/stills/rocket-engine_13.5.jpg: SHA-256 e739e273a15484b02a86d911a29b919d02b5123c8e1ce14c4a2af2e80b4a396d - film/stills/rocket-engine_18.6.jpg: SHA-256 80d65848c19790d3b8a871e422c4ddc73212fd30eff155cb0c8241283455c5f3 - film/stills/rocket-engine_23.4.jpg: SHA-256 a139b6550b653fb26d0eb103999e7853ff3c225dbd72236bbb4ea2640be60b0e - film/stills/rocket-engine_5.0.jpg: SHA-256 3443fc81a1e13aa25e6d95756e1bcfd6116c269d0ab9bc876d3a814c6e6dc309 - film/stills/rocket-engine_9.8.jpg: SHA-256 90898ad5cb04ca694010fe8adcc1f8c8b2320531950a65e218ef3838e9b0c5c3 - film/stills/turbofan_1.5.jpg: SHA-256 f4bcf30ab783ba36829131c61e6d09a0e6a0c31528c9149b11aaec1f0909d99a - film/stills/turbofan_13.5.jpg: SHA-256 8e999fca7a156376fca9fa45c1f7c283cadc10110814c42f1c710f22b154ab52 - film/stills/turbofan_18.6.jpg: SHA-256 341f928a65ee1462aff8c0e070836d1b4599022053b722e3c4043b94dcae1c37 - film/stills/turbofan_23.4.jpg: SHA-256 a26cc0ca784630ed8b8221f7c3bf303996d220cc6110873c2d08b19e814e5800 - film/stills/turbofan_5.0.jpg: SHA-256 9333c79658f74318dcb4f7c726e2a087b78e35f331b2afe239ea81ca83f2ef81 - film/stills/turbofan_9.8.jpg: SHA-256 a105a2bb8beaf30e1f649cdc0afb8ece56ff09b4e1573fb457c9d392bcff6ed7 - kit/lab-kit.js: SHA-256 6b6bf13e2f146eb725547538ce9700b0926b63885528cf8ed5f7b85cadb7f236 - kit/template.html: SHA-256 ec10cfb01f1b673e2f27ab7fec1921b85d415169edd241a30b008d41b385e26d - package.json: SHA-256 21e6afaca6fe0bd92cdd1428652d5b1a3aa57e200a236634b500adcc91f37d71 - references/solvers.md: SHA-256 9c40c1c0d88e4e7df67ce084e71f23e8aa86ff9434a2fdcb20f02a1612e5e2ec - scripts/inline.mjs: SHA-256 b3f38b47b79c30df49dc9622ce0014c63e0338299e3ce669030f745e086d0d2e - scripts/lib/browser.mjs: SHA-256 cf86daa5652502fd6bf73ff00fe4cff1d75690c756744456135d2c66c36b535d - scripts/lib/guard.mjs: SHA-256 06057c6011cb57cdfe711524987fb9587f99ba81f203b64e61a96d65962ce090 - scripts/lib/png.mjs: SHA-256 070beb972dfca9e34600c3bd03e487dfd4b71258d145db9a0f47e76fbc9b7cff - scripts/lib/serve.mjs: SHA-256 924e5deeaa4eb28e09f44226c7fdcd162735563f240c67e72fe43293d153cd33 - scripts/pinkscan.mjs: SHA-256 20796066ba4f4bf69bbdd94eb7555bd948bd94c554ca682b45d7420cb4cc2098 - scripts/record.mjs: SHA-256 fa7377c00e7187bc11f900cd181b5ae32c191383e17614afa1fdaf3682be2722 - scripts/shot.mjs: SHA-256 affdf0584c66d1bdeb2d4bdac2ea635fb4d51194a29f43fb514aac84cfb4efab If a file does not match, it changed since the listing was prepared: do not install it yet, tell me what you found and let me decide. 2. Read SKILL.md in full, skim the kit and the reference file, and read the scripts in scripts/, then tell me in plain words: what the skill does when used, which services it connects to, whether it reads or sends any personal information or keys, and anything that seems unusual. Stop on any red flag and tell me before going further. Confirm these points from the files: - The package is the whole repository, 53 files (about 11 MB, most of it example and film imagery): SKILL.md (the method: research the real machine, build from the kit template, geometry, materials, the solver, modes, critic rounds, the film), kit/ (lab-kit.js, the runtime that gets inlined into each generated page, and template.html), references/solvers.md, scripts/ (inline.mjs, which pastes the kit into a generated page; shot.mjs, the review screenshot script; record.mjs, the film recorder; pinkscan.mjs, which scans a finished film for stalled-GPU magenta frames; and lib/ helpers including a local file server), examples/ (four finished machines as single HTML files with their posters, the reference material to read before building), film/ (the launch film page and its stills), evals/evals.json, package.json (which declares playwright as a dev dependency for the scripts) and the project documents. The scripts are plain readable Node; nothing is obfuscated. - Nothing is sent anywhere during installation itself; the only download at install time is the repository from GitHub at the pinned commit. The files name no account, no API key and no telemetry, and the skill uploads nothing at use time: the simulation and the film are files on my machine. - Two use-time network touches to know about, both visible in the files: a generated page is a single HTML file with the kit inlined, but it still loads three.js ([email protected] and its addons) from cdn.jsdelivr.net and its fonts from Google Fonts (fonts.googleapis.com and fonts.gstatic.com) when the page is opened, so a generated page needs internet to display; and the film recorder's music option, --music freepd:<Category>/<Name>, downloads that track as an MP3 from archive.org (FreePD, CC0) at record time, while a local file or a URL I give can be used instead. The review and record scripts otherwise drive a local Chromium against the local page (served by scripts/lib/serve.mjs), and the research step may use the web, when available, for the real machine's dimensions and physics. - Filming needs Node 18 or newer, Playwright with Chromium and ffmpeg, as SKILL.md states; building a page needs Node only for the inline step. The skill's bar is that every number on screen is computed by the solver from real dimensions and equations, never typed in, and it requires several screenshot review rounds before a machine counts as done. 3. When I say "yes, install", install it as a skill: copy the whole repository folder (everything except .git) into the agent's skills directory, in a folder named "machine", keeping all 53 files. The upstream route, npx skills add AirsupHQ/machine-skill -g, installs the same repository; either route is fine as long as every file matches the fingerprints. 4. Verify with no network calls: frontmatter valid, all 53 files in place. 5. Report what was installed, where, and what I still need to do myself (for example: have Node 18+, Playwright with Chromium and ffmpeg ready before filming, and name a machine myself; using the skill is a separate step that only starts when I ask for it, and the use-time touches disclosed above happen only then). GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. If anything looks off, stop and ask me.