← Back to list
qisi-video-remix Checked
根据参考视频创意改编新故事，直接展示可读剧本、人物造型、分镜表和可独立使用的分段视频提示词。适用于参考视频改编、同款叙事创作和 AI 视频制作前的策划。
8.7Overall rating
Installation
🤖 Install via AI
Copy the prompt below and send it to your AI assistant (e.g. Claude Code) — it will follow the instructions and install automatically.
Install the "qisi-video-remix" skill by following the instructions at https://skill123.me/install/qisi-video-remix.
⌨️ Command line install
Run in your terminal — downloads and installs to ~/.claude/skills/.
curl -fsSL https://skill123.me/install/qisi-video-remix.sh | bash
📦 Download ZIP
Download the zip and extract it into your skills directory (e.g. ~/.claude/skills/), then restart your session.
⬇ Download v1.0.0 · 18 KBAbout this skill
Overview
根据参考视频创意改编新故事，直接展示可读剧本、人物造型、分镜表和可独立使用的分段视频提示词。适用于参考视频改编、同款叙事创作和 AI 视频制作前的策划。
Source
- Repo: https://github.com/wyuzhi/qisi-video-remix
- Path: (repo root)
Score breakdown
Trigger
6.5
Clear Chinese What/When for adaptation-style planning; description lacks an explicit When-Not boundary (only SKILL.md body limits scope to planning), and trigger vocabulary is moderate.
Structure
8.0
96-line SKILL.md with one delivery-format reference and one full worked example; clean and minimal, but ships a repo README.
Workflow
9.2
Strong execution discipline: mandatory user choice of 15s/30s segment cap with no default, declared defaults table, evidence-honesty rules for partial material, a concrete segmentation algorithm, self-contained-prompt rule, and a pre-delivery checklist with duration arithmetic checks.
Content
9.0
Precise, consistent imperative writing with good anti-fabrication rules; no commands to validate by design.
Engineering
10.0
Valid frontmatter; no scripts by design; both referenced files exist; openai.yaml interface file present.
Security
10.0
Capability manifest matches declared scope: no scripts, no network calls, no credentials; reference materials explicitly treated as data-not-instructions (anti prompt-injection); honest handling of unknown video metadata.
Pure planning skill (no scripts, no network): turns a reference video's narrative method into a new story with script, character/scene image prompts, storyboard and self-contained per-segment video prompts. Capability manifest (6.0): reads only user-supplied video/paths/links/screenshots/transcripts; sole side effect is optionally writing creative-plan.md (declared); no network destinations (explicitly provides no downloader, transcription model or generation service); no credentials; no persistence. Capabilities match declared scope, with an explicit data-not-instructions rule for attached video/subtitle documents.