技能详情（站内镜像，无评论）
许可证：MIT-0
MIT-0 ·免费使用、修改和重新分发。无需归因。
版本：v0.2.1
统计：⭐ 0 · 43 · 0 current installs · 0 all-time installs
⭐ 0
安装量（当前） 0
🛡 VirusTotal ：良性 · OpenClaw ：可疑
Package：afengzi/davinci-auto-editor
安全扫描（ClawHub）
- VirusTotal ：良性
- OpenClaw ：可疑
OpenClaw 评估
The skill does what it says (scan local media, call a cloud editing API, and generate EDLs) but it will upload local file metadata (including absolute paths) to a remote API and the example points at a raw IP address — this raises privacy and trust concerns you should review before using.
目的
Name, description, manifest, SKILL.md, README, and scripts/index.js are consistent: the skill scans a material folder, builds a materials index, calls a cloud API for a plan, and writes Resolve-importable files locally. Requiring only Node is proportionate to the described functionality.
说明范围
Runtime instructions and the implementation explicitly instruct scanning the entire material directory and POSTing a materials index to the cloud. The materials payload includes absolutePath, relativePath, name, size, modifiedAt and a SHA1-based id (path hashed). Sending absolute paths and file metadata to a remote API is outside pure local editing work and is a privacy/exfiltration risk unless you trust the target API.
安装机制
No install spec is provided (instruction-only plus included Node script). That minimizes installer risk. The project ships a local Node script (scripts/index.js) which will run on the user's machine — review that script (included) before running, but there is no remote installer download of code at runtime.
证书
The registry metadata lists no required env vars, but the config schema and examples require api_base_url and api_key (supplied in a config file). Requiring an API key is expected for a cloud-driven skill, but the example config uses a raw IP address (http://43.137.46.105:8787) which is unusual and potentially suspicious. Also, secrets are expected to be placed in a JSON config file (not an env var), which could leave credentials on disk if no…
持久
The skill does not request always:true, does not modify other skills, and only writes outputs under a project-adjacent _davinci_auto_editor/<taskId> directory. It does create files (resolve-import.json, timeline.edl, execution-report.json) in the filesystem as part of normal operation, which is expected for its purpose.
scripts/index.js:14
Environment variable access combined with network send.
examples/config.example.json:2
Install source points to URL shortener or raw IP.
scripts/index.js:114
File read combined with network send (possible exfiltration).
安装（复制给龙虾 AI）
将下方整段复制到龙虾中文库对话中，由龙虾按 SKILL.md 完成安装。
请把本段交给龙虾中文库（龙虾 AI）执行：为本机安装 OpenClaw 技能「DaVinci Auto Editor」。简介：Generate DaVinci Resolve import packages from local media plus a cloud editing …。
请 fetch 以下地址读取 SKILL.md 并按文档完成安装：https://raw.githubusercontent.com/openclaw/skills/refs/heads/main/skills/afengzi/davinci-auto-editor/SKILL.md
（来源：yingzhi8.cn 技能库）SKILL.md
---
name: davinci-auto-editor
description: Generate DaVinci Resolve import packages from local media plus a cloud editing API. Use when Codex needs to scan a material folder, request a cloud editing plan, and write a Resolve-importable EDL package with pure Node on the user machine.
metadata: {"openclaw":{"emoji":"🎞️","homepage":"https://github.com/imfengziaaa/video-auto-editor-skills","requires":{"bins":["node"]},"skillKey":"davinci-auto-editor","os":["darwin","linux","win32"]}}
---
# DaVinci Auto Editor
使用这个 skill 时，按下面顺序执行：
1. 读取 `examples/config.example.json` 同结构的配置文件。
2. 校验 `api_base_url`、`api_key`、`material_path`、`timeline_fps` 等关键参数。
3. 递归扫描素材目录，并向云端上报素材索引。
4. 调用云端 API 创建任务并获取剪辑计划。
5. 在本地只生成最小执行计划，不要把完整云端内部逻辑写入本地文件。
6. 由 Node 生成 Resolve 可导入的 `timeline.edl` 和导入说明文件。
7. 将准备结果回传云端。
## 输入参数
至少提供这些字段：
- `api_base_url`
- `api_key`
- `project_type`
- `aspect_ratio`
- `material_path`
- `template_id`
- `subtitle_mode`
- `music_policy`
- `pace_policy`
- `output_mode`
- `render_preset`
- `timeline_fps`
- `timeline_resolution`
可选字段：
- `task_timeout_ms`
- `poll_interval_ms`
- `request_timeout_ms`
- `task_name`
- `webhook_url`
- `extra_metadata`
## 输出结果
默认在素材目录旁创建 `_davinci_auto_editor/<taskId>/`，包含：
- `resolve-import.json`：最小本地导入计划
- `timeline.edl`：Resolve 导入文件
- `IMPORT-TO-RESOLVE.txt`：导入说明
- `execution-report.json`：本地执行报告
## 推荐工作流
- 把核心决策、模板逻辑、API Key 鉴权和配额管理放在云端服务。
- 本地只保留素材扫描、结果导出和回传逻辑。
- 优先使用短路径、稳定命名的素材目录，减少 EDL relink 成本。
- 在正式任务前先用样本素材验证时间线 FPS 和素材命名。
## 依赖要求
- Node.js 18 或更高版本
- 已安装 DaVinci Resolve
- 可访问云端 API 的网络环境
## 错误处理原则
- 缺少配置、素材目录不存在、API 调用失败时立即停止并返回非 0 退出码。
- 始终写出 `execution-report.json` 以便排查。
- 不在本地输出完整云端推理结果，只输出导入所需的最小执行数据。
- 明确提示第一版只覆盖基础拼接和时间线导入准备。