podcast-writer
根据用户给出的文字或者用户指定的txt/md文件内容，从类型列表中选择合适的文本类型，按照指定的格式给出podcast直播口述稿和音频文件
When & Why to Use This Skill
The podcast-writer skill is an end-to-end automation tool designed to convert raw text, documents, or web links into professional podcast scripts and high-quality audio files. It intelligently categorizes content—ranging from news analysis to technical deep-dives—and applies specialized templates to ensure the output is optimized for natural oral delivery, significantly reducing the time required for media production and content repurposing.
Use Cases
- Content Repurposing: Automatically transform long-form blog posts, articles, or research papers into engaging podcast scripts for multi-channel distribution.
- Automated News Briefing: Convert daily news links or RSS feed content into conversational scripts and audio files for morning briefings or social media updates.
- Educational Audio Creation: Turn technical documentation or study notes into easy-to-listen-to audio guides, making complex information more accessible on the go.
- Professional Interview Prep: Generate structured interview scripts based on guest biographies or specific topics using specialized 'Famous Interview' templates.
| name | podcast-writer |
|---|---|
| description | 根据用户给出的文字或者用户指定的txt/md文件内容，从类型列表中选择合适的文本类型，按照指定的格式给出podcast直播口述稿和音频文件 |
| allowed-tools | Read, Grep, Glob, Write, Search |
podcast-writer
Instructions
首先从环境变量root_dir中获取当前项目的路径，然后： 1、如果用户指定了输入的txt或者md文件，那么从文件中读取输入内容{input} 2、如果用户给出一个链接，调用工具从链接中读取输入内容{input} 3、如果用户直接给出了内容，则这个内容就是输入内容{input} 4、对输入内容进行分析，判断这个内容适合使用什么样的方式来输出podcast直播口述稿，可选的类型列表如下： ["名人访谈录","中美贸易战","新闻分析","新技术","其他"] 所有无法找到明确分类的都归类为"其他" "名人访谈录"对应的提示词模版为: FAMOUS_TEMPL.md "中美贸易战"对应的提示词模版为: TRADE.md "新闻分析"对应的提示词模版为: NEWS.md "新技术"对应的提示词模版为: TECH.md "其他"对应的提示词模版为: OTHER.md 5、为这个内容生成一个合适的标题{title}
文本输出
6、根据给出的提示词模版和输入{input},直接生成podcast直播口述稿，并写入当前项目目录下的output/{title}{currentdate_currenttime}.txt 7、将{input}单独写入当前项目目录下的/output/{title}{currentdate_currenttime}original.txt 8、第6步生成的文件内容还是很拗口，请将其改成更口语化的表达，除了口述稿正文内容，不要添加其他任何标题、旁白、解释,并将结果写入当前项目目录下的/output/{title}{currentdate_currenttime}_final.txt
生成音频文件
运行帮助脚本runninghub_api.py,使用当前项目目录下的/output/{title}_{currentdate_currenttime}final.txt 作为input.txt,使用当前项目目录下的/output/{title}{currentdate_currenttime}_final.flac 作为output.flac
# 使用指定文本文件和指定输出路径
python scripts/runninghub_api.py man input.txt output.flac