podcast-workflow_skill

Overview

This skill implements a one-stop podcast processing workflow from YouTube to Feishu knowledge base. It supports two entry points: periodic update checks for subscribed channels and direct processing from a YouTube link. The pipeline extracts captions, generates a content-digest (summary, key points, highlights), saves a local Markdown file, and auto-publishes to Feishu.

How this skill works

When triggered it either fetches recent updates from followed channels and presents a selection list, or accepts a direct YouTube URL. The skill runs a subtitle extractor, asks the user to confirm content-digest processing, produces a formatted summary (title, 10–15 key points, and highlight snippets), saves the Markdown locally, and calls a Feishu save script to publish the document. Optional step: convert the condensed key-points into a shareable image poster via a markdown-to-image flow.

When to use it

• You want an automated pipeline from YouTube episodes to a Feishu knowledge article.
• You want to periodically check subscribed channels for new podcast episodes and process selected items.
• You have a specific YouTube podcast URL and want a ready-made summary and highlights saved/published.
• You need consistent output format for daily podcast recommendations and easy sharing to social platforms.
• You want optional image posters generated from concise key-point summaries for social media.

Best practices

• Use the update-check entry for regular monitoring (recommended interval: 2 days) to avoid missing new episodes.
• Always confirm content-digest processing after subtitle extraction to allow manual skips or edits.
• Maintain the local save path and Feishu parent token in configuration to ensure reliable saves.
• Limit the poster conversion to concise versions (≤500 characters after trimming) and add emoji numbering for readability.
• Handle failures by prompting the user: offer manual transcript upload if subtitle extraction fails and check Feishu permissions on write errors.

Example use cases

• Run 'Get podcast updates' to list new episodes from tracked channels, pick several, and batch-process them into the knowledge base.
• Provide a single YouTube link and get an auto-generated Markdown file saved locally and published to the 'Daily Podcast Recommendations' Feishu space.
• Transform the top 8 key points into a PNG poster for sharing on social platforms using the markdown-to-image step.
• Recover from subtitle-extraction failure by uploading a transcript manually and then running the content-digest step.
• Use the pipeline as a daily content source for internal reading lists and newsletter highlights.