Produce Chinese video recaps with video-recap-skills
Use video-recap-skills when an agent should turn source video into a Chinese narration recap with scene understanding, scripting, voiceover, subtitles, and an optional editable JianYing draft.
npx skills add agentskillexchange/skills --skill produce-chinese-video-recaps-with-video-recap-skills
Use video-recap-skills when an operator wants an agent-hosted production workflow for video recaps instead of manually coordinating analysis, scripting, narration, and rendering tools. The operator installs the skill package in Claude Code, Codex CLI, OpenCode, or OpenClaw, provides source video paths and story context, and lets the video-recap skill orchestrate scene detection, ASR, VLM analysis, story planning, clip selection, narration writing, MiMo or Fish Audio voiceover, subtitle rendering, ffmpeg assembly, and optional JianYing draft export. Invoke this when the useful outcome is a repeatable agent workflow that produces a recap artifact from local media while preserving checkpoints, reusable analysis, advisory review, and editable outputs. The scope boundary is supervised video-recap production through the repo’s six declared skills; it is not a generic video editor, a raw ffmpeg wrapper, a Xiaomi MiMo listing, or a broad agent-plugin marketplace card.
What this skill actually does
Inputs and prerequisites: Claude Code, Codex CLI, OpenCode, or OpenClaw; Python 3.10+; ffmpeg with libass/subtitles support; Xiaomi MiMo API key; optional Fish Audio API key; local source video files; optional JianYing/CapCut-compatible draft workflow.
Setup notes: For Claude Code, run /plugin marketplace add zenstory-ai/video-recap-skills and /plugin install video-recap-skills@video-recap. For Codex CLI, run codex plugin marketplace add zenstory-ai/video-recap-skills and codex plugin add video-recap-skills@video-recap. Install Python 3.10+, ffmpeg, set MIMO_API_KEY, then ask the agent to check the video-recap environment before producing a recap.
Source and verification boundary: use https://github.com/zenstory-ai/video-recap-skills/blob/main/README.en.md as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.