Zyt one click video creation

用户输入一个选题或口播稿，自动生成完整短视频成片（文案、分镜、数字人口播 + AI 画面混剪）。适用于「一键成片」「根据选题做视频」等场景。当前 skill 已内聚 TTS、数字人视频生成与 AI 文生视频调用能力。

This skill is internally coherent for one‑click video creation, but before installing: (1) verify you trust the chanjing API domain (default https://open-api.chanjing.cc), because text, TTS requests, audio uploads and video prompts are sent there; (2) be aware the skill reads/writes ~/.chanjing/credentials.json (or $CHANJING_CONFIG_DIR/credentials.json) and will persist app_id/secret_key and access_token — do not put highly privileged secrets there unless you trust the service; (3) ensure ffmpeg/ffprobe are installed and sandbox the skill if you want to limit file or network access; (4) inspect the workflow.json you pass to run_render.py so it does not contain secrets or data you don't want transmitted. If you need stronger guarantees, ask the maintainer for an explicit credential storage policy or use a dedicated least-privilege API key/account for this skill.

Type: OpenClaw Skill Name: zyt-one-click-video-creation Version: 0.1.0 The skill bundle provides a legitimate and well-documented workflow for automated video creation, integrating Text-to-Speech, digital human avatars, and AI video generation via the Chanjing (蝉镜) API. The orchestration script (run_render.py) uses ffmpeg and ffprobe for local media processing, while the Python clients in the clients/ directory handle standard API interactions and authentication using a local credentials file (~/.chanjing/credentials.json). The bundle includes extensive Markdown templates (SKILL.md, storyboard_prompt.md, etc.) that provide high-quality instructions for the AI agent to ensure content consistency and avoid generation artifacts. No indicators of data exfiltration, malicious execution, or unauthorized persistence were found.