Platform Spec Export

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, and nothing else to install first. For using the skill afterwards, read this first: its only bundled script, scripts/export_platform.py, is local-only and needs Python 3 plus the ffmpeg and ffprobe command-line tools on PATH (the script calls them locally to probe the master video and to re-encode it), and Pillow (PIL) only for the optional --safe-preview overlay (imported inside that one function only). Pillow is not part of the Python standard library and is not installed by this install; the repository root requirements.txt lists pillow>=10 (pertinent for the safe preview) and numpy>=1.24 and edge-tts>=6.1 (neither is ever imported by this skill — they belong to other skills in the repository). No generation API, no key and no online link are used by this skill at all. Install "platform-spec-export" for me. It converts one finished master video (SKILL.md recommends making it 1080×1920 at 30 fps) into platform-ready files in one run: nine presets — douyin, tiktok and tiktok-shop at 1080×1920, WeChat Channels (channels) at 1080×1920, xiaohongshu at 1080×1440 (3:4), taobao-main at 1080×1080 (1:1) and taobao-3x4 at 1080×1440, amazon at 1920×1080 (16:9), and a generic square at 1080×1080 — all encoded the same way (H.264 High profile, yuv420p, CRF 20 with a per-preset bitrate cap of 8M, 10M or 12M, AAC 160k at 48 kHz when the master has audio, +faststart), resizing by blurred fill that keeps the whole frame (--mode blur, the default) or by centre crop (--mode crop, with --focus-y, default 0.45). It writes one <master>_<platform>.mp4 per requested platform plus an export_report.json spec report (resolution, duration, size, and a note per platform), and it can overlay the empirical UI safe zones on one frame with --safe-preview (top status bar about 9%, right action rail about 15% wide, bottom caption/product-card area about 22%). SKILL.md stresses that the presets are conservative recommended encodes, not official platform caps: duration, file-size and other upload rules change often and must be confirmed in each platform's official upload page or seller centre before uploading. Repository: https://github.com/xianyu110/ecommerce-video-skills (the skill is the whole skills/platform-spec-export/ folder at the pinned commit: SKILL.md plus its scripts/ folder, 2 files) Pinned version: install the code at commit 96c49b349b7d975bda8f05ce6669c1d33500c269 exactly. Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Clone https://github.com/xianyu110/ecommerce-video-skills into a temporary folder, check out commit 96c49b349b7d975bda8f05ce6669c1d33500c269, and copy the skills/platform-spec-export/ folder from that checkout. Then compare every file in the folder with the fingerprints recorded for this listing on 2026-10-19: [FINGERPRINTS] - SKILL.md: SHA-256 746b696329e9c44ac37ee04eb2a749d29410a6439399fa14dc1df320e4cfedc2 - scripts/export_platform.py: SHA-256 988ac7e0cd775b4580dd6bfd37a1121cf37379979fe5bbd0a76b31902e41e234 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 and read the only script in scripts/ (export_platform.py), and 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 skill folder holds 2 files only: SKILL.md (3,940 bytes) and one Python script in scripts/ (export_platform.py, 6,320 bytes). That script is a byte-identical copy of the same-named script at the repository root (scripts/export_platform.py, same SHA-256): the skill is self-contained in its own folder and needs no file from elsewhere in the repository — not the other 12 skills, not the root scripts/, not the README, and not the repository requirements.txt at install time. SKILL.md only ever calls the script sitting next to it (scripts/export_platform.py, relative to the skill folder). Two mentions in SKILL.md are not calls by this skill and nothing in this folder requires them: the advice to re-render a 9:16 master at 1920×1080 with assemble.py instead of transcoding it when 16:9 matters (assemble.py is not in this folder — it belongs to the repository root scripts/ and to another skill of the repository), and the checklist pointer to the cover-title-ab skill for making separate 3:4 covers for WeChat Channels / Xiaohongshu. - Nothing is sent anywhere during installation itself; the only download at install time is the skill folder from GitHub at the pinned commit. No script is run during installation, no key is asked for, and there is no telemetry in the folder. - What leaves this machine at use time, disclosed here before it ever happens: nothing. Using the skill is a separate step that only starts when I ask for platform exports, and it sends nothing to anyone: scripts/export_platform.py makes no network call at all — it imports only argparse, json, os and subprocess at the top level (plus Pillow, from PIL, inside the safe_preview function only), contains no URL, no API client and no key handling, and SKILL.md defines no API-key route and no online-link route for this skill. The only URL in the whole folder is the plain GitHub link in SKILL.md to this repository itself, which nothing opens during installation or use. The prompt template in SKILL.md is text I use with you locally in our conversation. Uploading the exported files to Douyin / TikTok / TikTok Shop / WeChat Channels / Xiaohongshu / Taobao / Amazon or any other platform, and confirming each platform's current duration, size and content rules in its official upload page or seller centre, happen manually, by me — the skill itself never contacts any platform. - The local-only parts at use time: scripts/export_platform.py runs entirely on this machine — it calls the local ffprobe to read the master video's streams, duration and size, builds an ffmpeg filter locally (scale to the preset size as-is when the ratio already matches, a blurred-fill background plus the full frame on top in blur mode, or a scale-and-centre-crop in crop mode positioned by --focus-y), calls the local ffmpeg to re-encode each requested preset (libx264, High profile, medium preset, CRF 20, the preset's maxrate/bufsize, yuv420p, +faststart, AAC 160k 48 kHz stereo when audio is present and no audio track otherwise), re-probes each output locally with ffprobe, and writes the <master>_<platform>.mp4 files plus export_report.json to the local output folder (default out/export). With --safe-preview it extracts one frame at 1.0 s with the local ffmpeg, draws the three empirical safe zones on it locally with Pillow, saves the PNG and deletes its temporary frame file. The preset table, the safe-zone guidance, the prompt template and the checklist in SKILL.md also stay local, between you and me. Rules from SKILL.md: the presets are safe recommended encodes, not official caps — never invent a specific platform limit number; in crop mode check that subtitles and the product were not cut off; and for an important 16:9 platform a re-render at size is preferred over transcoding because transcoded subtitles get smaller. - Do not install Pillow, ffmpeg/ffprobe or anything else during this install, do not run export_platform.py, and do not export any video: setup of the use-time dependencies is my decision, later, when I actually ask for exports. 3. When I say "yes, install", install it as a skill: copy the whole platform-spec-export folder into the agent's skills directory, in a folder named "platform-spec-export", keeping all 2 files (SKILL.md at its root and the one script in its scripts/ subfolder). A manual clone followed by copying the skills/platform-spec-export folder into ~/.claude/skills/ is the route described here; any other route is fine only as long as every installed file matches its fingerprint. 4. Verify with no network calls: frontmatter valid (name: platform-spec-export), all 2 files in place with matching fingerprints. Do not run export_platform.py to verify. 5. Report what was installed, where, and what I still need to do myself (for example: provide one finished master video and choose the platforms, the blur-or-crop mode and the output folder myself; install Python 3 and have ffmpeg/ffprobe on PATH, and Pillow if I want the safe-zone preview, if I want the script to export; confirm each platform's current limits in its official upload page or seller centre and upload the files myself; using the skill is a separate step that only starts when I ask for it, and no use-time send happens on it at all, as disclosed above). 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.