# Social Live Photo Card > Turn user-provided short videos, video collections, photo collections, or mixed media into a social-platform Live Photo card when one small motion point explains more than a static image, with collection screening, first-frame, duration, crop, and publish-path checks. - **Type:** Skill - **Install:** `agentstack add skill-justinmchoi-ai-toolkit-social-live-photo-card` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [justinmchoi](https://agentstack.voostack.com/s/justinmchoi) - **Installs:** 0 - **Category:** [Content & Media](https://agentstack.voostack.com/c/content-and-media) - **Latest version:** 0.1.0 - **License:** MIT - **Upstream author:** [justinmchoi](https://github.com/justinmchoi) - **Source:** https://github.com/justinmchoi/ai-toolkit/tree/main/skills/media/social-live-photo-card ## Install ```sh agentstack add skill-justinmchoi-ai-toolkit-social-live-photo-card ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About # Social Live Photo Card Create a social card that behaves like a still image first and a short motion asset second. The output should still read as a publishable card when the motion does not play. Use this skill when the user has media for a social post and asks for a Live Photo, dynamic card, animated social card, Rednote/Xiaohongshu Live Photo, or WeChat article Live Photo. Accept a single file or a folder/collection of photos, videos, screen recordings, or mixed media. ## Inputs - User-provided video clips, screen recordings, product captures, short lifestyle clips, photo sets, or folders containing mixed photos and videos. - Optional text, article excerpt, product claim, tutorial step, or target platform. - Optional existing static card style, brand constraints, or visual examples. Default to user-provided media. Use public media only for explicitly approved demo or test work, and keep source attribution. Treat photo-only collections as synthetic-motion Live Photo candidates. They can be converted into a short pan, zoom, parallax-like crop, or slideshow motion asset, but they do not preserve real capture-time Live Photo movement. ## Output Write all generated artifacts into the task output folder, not into the skill directory: - static first-frame JPG or PNG - preview video for review - MOV or platform-required motion asset when tooling is available - `.pvt` package only when a local packager exists and is verified - collection contact sheet, candidate sheet, or selection note when more than one source file is provided - short delivery note explaining platform duration, publish path, and any remaining manual phone-side step ## Workflow ### 1. Confirm Platform And Motion Point Identify the target: - Rednote/Xiaohongshu: design around a `5s` Live Photo. - WeChat article: design around a `3s` Live Photo. - Local test or unknown platform: ask or choose the stricter `3s` default. Name the single motion point the viewer should understand without audio. If the clip needs a full tutorial, narration, or a long sequence, recommend static multi-card instructions or video instead of Live Photo. ### 2. Classify And Screen The Material Classify the source before editing: - Single video: choose the best 3-5 second time range. - Video collection: create per-video thumbnails or contact sheets, then choose one winning clip before detailed editing. - Photo collection: choose one hero still and define synthetic motion that keeps the card readable as a still. - Mixed collection: choose whether the hero still comes from a photo, a video frame, or a composed card; choose the motion asset from the strongest short video unless a synthetic photo motion is clearer. Prefer clips near 10 seconds or less. If the source is longer, unclear, or a folder with multiple candidates, create a sparse contact sheet of 8-15 frames or thumbnails before editing. Use it to check for: - black or empty first frames - transitions crossing pages or scenes - UI text too small for mobile - key result not yet visible - subject, face, product, button, or result area getting cropped - duplicate, weak, or off-topic collection items Ask the user to choose a time range when the contact sheet does not make the best segment obvious. ### 3. Choose The Card Structure Pick one structure: - Single video card: one strong clip, optionally with one short title group. - Two-, three-, or four-panel collage: multiple attractive clips where the visuals carry the message and text can stay minimal. - Triple Live Photo: three parallel results, states, views, or examples. Do not use this for a sequence that must be understood in order. - Synthetic-motion photo card: one or more photos animated with subtle pan, zoom, hold, or slideshow motion. Label it as synthetic motion in the delivery note. - No Live Photo: motion is decorative, confusing, too long, or less clear than a static card. ### 4. Make The First Frame Work As A Static Card Before generating motion assets, render or inspect the first frame as a normal card: - target the platform aspect ratio, usually `3:4` - keep the subject and result visible - avoid covering faces, products, buttons, or result areas with text - use one concise title group only when it improves comprehension - preserve a quiet zone for text instead of defaulting to full-image overlays - do not put production notes, keyboard shortcuts, or instructions on the card If the first frame fails, fix crop, segment, or layout before continuing. For collections, the first frame must also explain why the selected item won: prefer the clearest subject, strongest emotion/action, best product evidence, or most recognizable place over a merely busy frame. ### 5. Generate And Package Motion Assets Generate the shortest asset that proves the motion point. Keep the dynamic portion aligned with the static card crop and typography. When packaging is available, produce the platform package expected by the user's publishing flow. When `.pvt` packaging is not available, deliver the verified JPG/MOV pair and state the missing packaging step plainly. For photo-only synthetic motion, keep movement subtle enough that the first frame remains the primary card. Avoid aggressive slideshow cuts unless the user explicitly asks for a multi-photo card. ### 6. Verify Before Delivery Run a visual and file-level check: - first frame works as a standalone social card - duration matches the target platform (`3s` or `5s`) - action is understandable without audio - crop, safe areas, and text placement survive mobile viewing - MOV/preview opens locally - collection selection is documented when multiple source files were provided - package was written to the task output folder - delivery note explains that phone-side publishing may be required and desktop upload may not preserve Live Photo behavior ## Suitable Material - product updates and AI tool demos where a click or generated result matters - webpage, code, image, or UI generation proof - one game action, timing, route, or success moment - one tutorial step that needs motion evidence - travel, lifestyle, food, craft, and product-detail clips with natural motion - several parallel product states, model outputs, or detail angles Motion should provide evidence. If it only decorates the card, use a static card. ## Stop Conditions Stop or switch format when: - the user wants a long-form video, full tutorial, or edited reel - the source requires audio or narration to make sense - the best segment cannot be chosen without user judgment - the media is not user-owned or licensed for the requested use - platform packaging tooling is unavailable and the user requires a ready-to-post phone package - the output would need cloning upstream templates or assets with incompatible licensing ## Provenance This local skill is adapted from public methodology signals in an X post and GitHub README by `op7418`, but it does not import upstream files, templates, assets, scripts, or skill text. See `references/source-intake.md`. ## Source & license This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [justinmchoi](https://github.com/justinmchoi) - **Source:** [justinmchoi/ai-toolkit](https://github.com/justinmchoi/ai-toolkit) - **License:** MIT Install and usage instructions live in the source repository linked above. ## Pricing - **Free** — Free ## Security capabilities Automated source analysis of v0.1.0 — what this tool can access: - **Network access:** no - **Filesystem access:** no - **Shell / process execution:** no - **Environment & secrets:** no - **Dynamic code execution:** no *"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.* ## Versions - **0.1.0** — security scan: passed — Imported from the upstream source. ## Links - Listing page: https://agentstack.voostack.com/l/skill-justinmchoi-ai-toolkit-social-live-photo-card - Seller: https://agentstack.voostack.com/s/justinmchoi - Browse the marketplace: https://agentstack.voostack.com/browse --- Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.