Automated Content Pipeline from Podcasts and Video

★ 7.5 · content

content-pipeline automates the full content-production cycle in Claude Code — from mining quotable moments to gating platform-native drafts before publish. The pipeline chains five non-interactive Python scripts: quote-mining-engine scans podcast RSS feeds and meeting notes for viral-worthy candidates; editorial-brain runs a two-pass LLM analysis on video transcripts to find clip-worthy hook→build→payoff moments scored 0–100; content-transform repurposes long-form source into X threads, LinkedIn posts, YouTube Short scripts, and newsletter sections; content-quality-scorer evaluates batches on five heuristic dimensions (voice similarity, specificity, AI-slop penalty, length, engagement) without any API call; and content-quality-gate acts as a CI-style publish filter. Dependencies are anthropic and feedparser; video cutting additionally requires yt-dlp and ffmpeg. The skill reuses expert panels and rubrics from the sibling content-ops skill for its optional in-loop scoring step. It suits anyone needing to turn a podcast feed or transcript into scheduled social posts entirely without manual intervention.