Selling-Point Storyboard

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: it is a text-only planning skill with no bundled script at all — it works in our conversation, writing a storyboard table and a storyboard.json file locally. No Python package, no ffmpeg and no generation service is needed to produce the storyboard itself. SKILL.md only suggests, as an optional local check at use time, validating the JSON with "python -m json.tool storyboard.json" (Python 3 standard library only, nothing to install). Rendering the storyboard into a video is NOT part of this skill: SKILL.md points, for field definitions and for that separate rendering step, to the sibling skill ffmpeg-auto-assemble and its script skills/ffmpeg-auto-assemble/scripts/assemble.py, which are not in this skill's folder, are not installed by this install, and must not be run here. No generation API, no key and no online link are used by this skill at all. Install "selling-point-storyboard" for me. It turns a product's real selling points into a shot-by-shot storyboard for a 15, 30 or 60 second ecommerce short video: sort the selling points (strongest, easiest to prove on camera, in shots 2–3), use the structure formula for the chosen length (15s: hook, product reveal, 2–3 selling points, call to action, 5–6 shots; 30s: hook, pain point, reveal, 3 selling points each "say it + prove it", use scene, call to action, 8–10 shots; 60s: hook, personal experience, unboxing, 3–4 selling-point demos, comparison, downsides / who it is for, call to action, 14–18 shots), keep one piece of information per shot, pair every selling point with proof footage (demo, detail or comparison), and output two things: a Markdown table per shot (timecode, visual / source material, voice-over line of at most 2 sentences, subtitle with key words in **bold**, optional pop-up title text, camera motion) and a render-ready storyboard.json (top-level voice, rate, bgm and tag, plus shots with image or video, motion, line, sub, title, fit, start, min and cta). It checks the timing from the script length (Chinese about 4–5 characters per second, lines divided by 4.5; English about 2.5 words per second in the checklist), marks shots whose material is missing as TODO with a shoot-or-generate suggestion, and forbids inventing data, certifications or prices. Repository: https://github.com/xianyu110/ecommerce-video-skills (the skill is the whole skills/selling-point-storyboard/ folder at the pinned commit: SKILL.md only, 1 file, no scripts/ folder) 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/selling-point-storyboard/ 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 7f108e8d25533208e90b5fbb5b5b5d9a7800481f0a41424dcfcf531b80d65320 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 (it is the only file in the folder: there is no script to read), 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 file: - The skill folder holds 1 file only: SKILL.md (5,019 bytes), with no scripts/ folder and no other file. Unlike the sibling cover-title-ab skill, there is no bundled script here at all. SKILL.md is self-contained for its own job — writing the storyboard table and the storyboard.json — and needs no file from elsewhere in the repository to do it: not the other 12 skills, not the root scripts/, not the README, and not the repository requirements.txt (which lists pillow, numpy and edge-tts for other skills in the repository and is never used by this skill). The names it mentions are pointers for later, separate steps only (see the next points), not files this install copies or requires. - 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 a storyboard, and it sends nothing to anyone: the folder contains no code at all — no imports, no URL-driven call, no API client and no key handling — and SKILL.md defines no API-key route and no online-link route for this skill. The director prompt template in SKILL.md (fill in duration, product, audience, platform, the real selling points and the available material files) is text I use with you locally in our conversation. The "voice" value in the storyboard.json template (for example "zh-CN-YunxiNeural") is only a text field written into that local JSON file for a later renderer; writing it contacts no voice service. - The cross-references in SKILL.md are not sends by this skill and are not installed here: SKILL.md says the storyboard.json field definitions live in the sibling skill ffmpeg-auto-assemble and gives that sibling's rendering command (python skills/ffmpeg-auto-assemble/scripts/assemble.py storyboard.json -o out/video.mp4) for if I later choose to render the video — that is a separate step with a separate skill, not part of this skill and not run here; its checklist points to hook-3s-script for the first-shot hook, and for a shot whose material is missing it suggests handing the shot to image-to-video-shots or to the sister repository ecommerce-image-skills (a plain GitHub link) to make the image — again suggestions for me, in text, that nothing opens or executes during installation or while writing a storyboard. - The local-only parts at use time: everything this skill does stays on this machine, between you and me — sorting my selling points, applying the duration structure, writing each shot's visual, voice line, subtitle and camera motion, checking the total timing against the line lengths, marking missing material TODO, and writing the Markdown table and the storyboard.json locally. The only command SKILL.md suggests at use time is the optional local JSON check "python -m json.tool storyboard.json"; do not run even that during this install. Rules from SKILL.md: one piece of information per shot, every selling point paired with proof footage, the first shot is the hook and the last is the call to action, and never invent data, certifications or prices — selling points must be real. - Do not install anything else during this install, do not install or copy ffmpeg-auto-assemble, hook-3s-script, image-to-video-shots or any other sibling skill, do not run assemble.py, and do not write any storyboard or render any video: using the skill is my decision, later, when I actually ask for a storyboard. 3. When I say "yes, install", install it as a skill: copy the whole selling-point-storyboard folder into the agent's skills directory, in a folder named "selling-point-storyboard", keeping its 1 file (SKILL.md at its root). A manual clone followed by copying the skills/selling-point-storyboard 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: selling-point-storyboard), the 1 file in place with matching fingerprint. Do not run any command from SKILL.md to verify. 5. Report what was installed, where, and what I still need to do myself (for example: provide the product information and 3 to 5 real selling points, the target duration of 15, 30 or 60 seconds, the list of material files I already have, and the platform, aspect ratio and voice language or tone myself; using the skill is a separate step that only starts when I ask for it, no use-time send happens on it at all as disclosed above, and rendering the storyboard.json into a video, or generating a missing image, would be a separate step with other skills that this install does not set up). 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.