Stream Helper Skills

A dynamic, skill-based alternative to the Stream Helper web app. Instead of clicking through a rigid multi-page UI, each production task (topic ideas, titles, thumbnails, transcription, ...) is a Claude Code skill you invoke directly, in whatever order suits the moment. Every skill writes its result to a numbered, human-readable file inside a per-project folder under projects/.

Install via npm

Add these skills to an existing project with one command:

npx stream-helper-skills [target-dir]

This copies .claude/skills/, scripts/, references/, and .env.example into target-dir (defaults to the current directory). Re-running it later skips files that already exist (so it won't clobber edits you've made) unless you pass --force.

Alternatively, clone this repo directly and open it in Claude Code — both are equivalent ways to get the same skills.

Setup

1. Node.js ≥20.12 on PATH. That's it — this package has zero npm dependencies.
2. ffmpeg and yt-dlp on PATH (only needed for /transcribe).
3. Copy .env.example to .env and fill in your OpenAI credentials:

   cp .env.example .env

Usage

Open this repo in Claude Code and invoke any skill, optionally passing a project name:

/topic-ideas my-spring-ai-episode
/youtube-titles my-spring-ai-episode
/transcribe my-spring-ai-episode --file ./recording.mp4
/chapters my-spring-ai-episode

If you omit the project name, Claude will infer it from your current directory (if you're inside projects/<name>/) or ask you.

A project's first skill run creates projects/<name>/00_project.md — fill in host/guest names, brand notes, and preferences there; every skill reads it for context.

Output files

Each task writes to one file, prefixed with its workflow-stage number:

| Stage | Files | |---|---| | 1 Plan | 01_plan_topic-ideas.md, 01_plan_guest-ideas.md | | 2 Titles | 02_titles_youtube-titles.md, 02_titles_youtube-description.md, 02_titles_youtube-tags.md | | 3 Thumbnail | 03_thumbnail_ideas.md, 03_thumbnail_prompts.md, 03_thumbnail_image.png (+ .md sidecar) | | 4 Announce | 04_announce_linkedin.md, 04_announce_social.md, 04_announce_hashtags.md | | 5 Transcribe | 05_transcription_transcript.md | | 6 Wrap-up | 06_wrapup_chapters.md, 06_wrapup_summary.md |

See references/file-naming.md for details, references/context-map.md for what each skill reads before generating, and references/validation-rules.md for the formatting rules applied to tags/hashtags/social posts/chapters.

Regenerating

Re-running a skill overwrites its output file with a fresh result. To keep a prior version, just ask Claude to refine or append instead — it's a normal file edit, no special skill needed.