KD(快导)

v1.0.4

快导(KD) - 短视频脚本批量生成与管理。当用户需要批量生成多平台短视频脚本、管理文案库时使用。

1· 62·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for gitgily/kuaidao.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "KD(快导)" (gitgily/kuaidao) from ClawHub.
Skill page: https://clawhub.ai/gitgily/kuaidao
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install kuaidao

ClawHub CLI

Package manager switcher

npx clawhub@latest install kuaidao
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
medium confidence
Purpose & Capability
The code, prompts and config clearly implement a multi-step short‑video script generator that reads/writes Excel, loads templates, and can optionally upload reports to Feishu. Those capabilities match the skill name/description. Minor inconsistency: registry metadata at the top lists no required binaries/env, while SKILL.md declares 'python' and pypi package 'openpyxl' — that is expected for the functionality but the registry summary omitted them.
Instruction Scope
SKILL.md and the prompt files instruct the agent to read local config files, read/write Excel workbooks (openpyxl), use a 'web_search' skill to collect trending titles, and optionally post reports to Feishu via lark‑cli. All referenced files/variables (platforms.json, user_config.json, Excel paths, optional LARK_CLI_TOKEN) are directly relevant to the declared tasks. The instructions do not ask for unrelated system credentials or to exfiltrate arbitrary files.
Install Mechanism
No install spec that downloads arbitrary code was provided; the package includes Python source files and requires openpyxl. There are no remote download URLs or archive extraction steps in the metadata. This is a local Python package style project — lower install risk than arbitrary remote installers.
Credentials
The skill requests no required environment variables and only one optional sensitive token (LARK_CLI_TOKEN) for uploading reports to Feishu. That optional credential is proportional to the documented Feishu upload feature. No unrelated secrets (AWS, GitHub tokens, etc.) are requested.
Persistence & Privilege
The skill is not set to always:true and does not declare modifications to other skills or global agent settings. It defines its own config paths under ~/.agents/skills/kd/, which is reasonable for a skill storing its templates and rules.
Assessment
This package appears to do what it says: generate and manage short‑video scripts, read/write Excel files, optionally search the web for trends, and optionally upload reports to Feishu. Before installing or running it: 1) Note the metadata mismatch — SKILL.md requires Python + openpyxl but the registry summary omitted them; ensure you have a safe Python environment and install openpyxl yourself. 2) Back up any Excel “文案库” files the skill will edit; the tool reads and writes user Excel files. 3) Only provide the optional LARK_CLI_TOKEN if you trust the skill — granting it allows writing/creating docs in your Feishu space (the config includes broad Feishu scopes). 4) Review config/platforms.json and config/user_config.json to ensure no unexpected paths or defaults are set (the skill will read local files under its config path). 5) Be aware the skill uses a web_search capability to collect trending data (network access) — if you want to avoid external queries, disable external_search or set external_keywords explicitly. If you want higher assurance, review the included scripts (especially feishu_permission_helper.py and any network‑related code in the scripts/ directory) or run the code in an isolated environment.

Like a lobster shell, security has layers — review code before you run it.

latestvk97af2jph9gsxq086s7vvmy5e585nm9v
62downloads
1stars
5versions
Updated 21h ago
v1.0.4
MIT-0

快导 (KD)

快导(KD)是一个跨平台、跨行业的短视频脚本批量生成与管理工具。

⚠️ CRITICAL — 首次使用必读

使用前必须完成以下配置:

  1. 安装依赖: pip install openpyxl
  2. 配置搜索关键词: 编辑 config/platforms.json
  3. 设置文案库路径: 编辑 config/user_config.json

详见 README.md 完整文档。

核心功能

功能说明
快导系列10步流程生成多平台短视频脚本
规则更新生成/更新平台运营规则文档

Shortcuts(快速使用参考)

用户说我执行说明
"快导小红书,生成5条"执行完整10步流程自动执行搜索→生成→写入→报告全流程
"快导第6步"执行单步生成脚本仅执行Step 6,需前置步骤数据
"配置关键词"引导配置 platforms.json设置平台搜索关键词池
"配置文案库路径"引导配置 user_config.json设置Excel文案库保存位置
"配置飞书空间"引导配置 report_space_id设置报告上传的飞书知识库
"检查文案库格式"执行Step 5格式扫描读取Excel实际格式参数
"查看平台规则"读取 rules 文档查看已生成的平台运营规则
"更新小红书规则"执行规则更新系列搜索最新规则并更新文档
"测试快导"运行 test_workflow.py快速测试核心功能是否正常

执行模式

自动模式(默认)

from scripts import WorkflowManager
workflow = WorkflowManager('xiaohongshu')
result = workflow.run_full()

交互模式(每步完成后暂停)

from scripts import WorkflowManager

# 启用交互模式
workflow = WorkflowManager('xiaohongshu', interactive=True)

# 执行,每步完成后返回等待
result = workflow.run_full()
if result.get('paused'):
    print(result['message'])  # "Step X 完成,是否继续?"

# 用户确认后继续
result = workflow.resume()

回调模式(即时回复每步结果)

def on_step_complete(step_num, result, should_pause):
    print(f"Step {step_num} 完成")  # 即时回复用户

workflow = WorkflowManager('xiaohongshu')
workflow.run_full(callback=on_step_complete)

核心模块

  • ConfigManager - 配置管理
  • ScriptGenerator - 脚本生成
  • ExcelManager - Excel操作

详细用法和示例见 README.md

版本

版本日期说明
1.1.02026-04-27新增:智能时长计算、活动管理、详细内容格式、BGM生成、Excel行高100
1.0.02026-04-24初始版本,支持3平台

v1.1.0 新功能详解

1. 智能时长计算

系统会自动计算并调整每个分镜的时间段:

  • 台词时长:字数 × 0.25秒
  • 运镜时长:固定2秒、推拉3秒、摇移4秒、复杂6秒、航拍8秒
  • 建议时长:max(台词, 运镜) + 1秒缓冲
  • 自动调整:使用建议时长更新B列时间段

配置config/platforms.json):

"duration_policy": {
  "xiaohongshu": {
    "auto_adjust": true,  // 开启自动调整
    "calculation_rules": {
      "speech_rate": 0.25,
      "motion_base": { "static": 2, "push_pull": 3, ... },
      "buffer_seconds": 1
    }
  }
}

2. 活动管理

支持自然语言管理活动:

  • 添加活动:添加活动:春节特惠,关键词:春节、过年
  • 列出活动:列出所有活动
  • 查询活动:查询活动:春节
  • 删除活动:删除活动:春节特惠

系统会根据脚本主题自动匹配活动,填入L列(关联活动)。

3. 详细内容格式

分镜内容现在支持更详细的描述:

  • C列-镜头:设备/焦段/内容
  • D列-运镜:方式/轨迹/速度
  • E列-技巧:光线/ISO/对焦/色调
  • F列-画面:场景/构图/色彩/氛围

4. BGM详细生成

I列BGM现在要求详细格式:

音乐名:《稻香》周杰伦
风格:田园治愈、轻快温暖
使用时机:0-30秒前奏,30-60秒主歌,60-90秒副歌

5. Excel行高统一

数据行统一设置为100,标题行20.4:

"row_height": {
  "title_row": 20.4,
  "data_row": 100
}

6. 外网搜索自动收集

user_config.json 新增 external_search 配置:

{
  "external_search": {
    "auto_collect": true,      // 空关键词时自动获取热门
    "collect_count": 20,
    "platforms": ["TikTok", "YouTube"],
    "default_keywords": []       // 备用关键词(可选)
  }
}

三种使用场景:

  1. external_keywords 有值 → 使用用户配置的关键词
  2. external_keywords 为空,default_keywords 有值 → 使用默认关键词
  3. 两者都为空 + auto_collect: true → 调用 web_search 自动获取当前热门

注意: 场景3如果搜索失败会直接报错,不使用备用数据。


联系我们

如有问题或建议,欢迎交流:

  • 微信号: huitouyoujianta
  • 用途: 快导(KD)技能使用咨询、功能建议、技术交流

欢迎加微信交流学习!

Comments

Loading comments...