Cover & Title A/B

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/cover_ab.py, is local-only and needs Python 3 with Pillow (PIL) and NumPy, plus the ffmpeg and ffprobe command-line tools on PATH (the script calls them locally to probe the video and extract frames), plus a CJK font — the script searches system font paths for Noto Sans CJK / PingFang / Microsoft YaHei Bold and exits with "no CJK font; pass --font" if it finds none. Pillow and NumPy are not part of the Python standard library and are not installed by this install; the repository root requirements.txt lists pillow>=10 and numpy>=1.24 (it also lists edge-tts>=6.1, which belongs to other skills in the repository and is never imported by this skill). No generation API, no key and no online link are used by this skill at all. Install "cover-title-ab" for me. It makes A/B cover and title variants for ecommerce short videos: write 6–10 short titles (each at most 14 characters in the SKILL.md guidance, covering different hook types — pain-point question, result, scene, number list, contrast, comparison), generate 3–4 cover versions in 3:4 or 9:16 by putting different titles and layouts on either clean still images I supply or the sharpest frames auto-picked from my finished video, export cover_A.jpg, cover_B.jpg and so on plus one ab_sheet.jpg comparison sheet, then design a one-variable-at-a-time A/B test and record exposure, click-through and completion in the table in SKILL.md. The script rotates four title styles (yellow with black stroke / white with red stroke / black on a yellow banner / white on a translucent black band). Repository: https://github.com/xianyu110/ecommerce-video-skills (the skill is the whole skills/cover-title-ab/ 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/cover-title-ab/ 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 edf9c36606f91f3701923c58dd28191bd341ce665b7f53ab21d384b89219d7ff - scripts/cover_ab.py: SHA-256 10b5b767f6a74f9f93942cfd5ee9d55de8fad261c0e1a3847b73d8a08b8bae67 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/ (cover_ab.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 (4,000 bytes) and one Python script in scripts/ (cover_ab.py, 7,140 bytes). That script is a byte-identical copy of the same-named script at the repository root (scripts/cover_ab.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/cover_ab.py, relative to the skill folder). - 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 covers or titles, and it sends nothing to anyone: scripts/cover_ab.py makes no network call at all — it imports only argparse, glob, os, subprocess, NumPy and Pillow, 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 title-writing prompt template in SKILL.md is text I use with you locally in our conversation. Two mentions in SKILL.md are not sends by the skill: the link to the sister repository ecommerce-image-skills (its xiaohongshu-cover skill, named only as an optional way to make a fancier cover base image) is a plain GitHub link that nothing opens during installation or use, and publishing the test and reading its click-through/completion numbers happen manually, by me, on the platform I choose (Douyin / WeChat Channels / Xiaohongshu / TikTok) or in that platform's own built-in cover/title test feature if it has one — the skill itself never contacts any platform. - The local-only parts at use time: scripts/cover_ab.py runs entirely on this machine — with --images it opens my clean stills locally with Pillow, otherwise it calls the local ffprobe to get the video duration and the local ffmpeg to extract candidate frames, scores their sharpness locally with NumPy (Laplacian variance), keeps the sharpest frames, draws the title text locally with the CJK font it finds (or the one I pass with --font), and writes cover_A.jpg, cover_B.jpg and so on plus ab_sheet.jpg to the local output folder (default out/covers). The title formulas, the prompt template, the A/B record table and the checklist in SKILL.md also stay local, between you and me. Rules from SKILL.md: one variable changed at a time, no absolute claims and no unprovable data in titles, and comparison titles only for things actually tested. - Do not install Pillow, NumPy, ffmpeg/ffprobe or anything else during this install, do not run cover_ab.py, and do not generate any cover: setup of the use-time dependencies is my decision, later, when I actually ask for covers. 3. When I say "yes, install", install it as a skill: copy the whole cover-title-ab folder into the agent's skills directory, in a folder named "cover-title-ab", 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/cover-title-ab 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: cover-title-ab), all 2 files in place with matching fingerprints. Do not run cover_ab.py to verify. 5. Report what was installed, where, and what I still need to do myself (for example: provide a finished video or clean still images and the hook/selling points, and choose the platform and ratio myself; install Python 3 with Pillow and NumPy and have ffmpeg/ffprobe on PATH, and a CJK font available or pass --font, if I want the script to render covers; 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.