process

A command that turns one or more YouTube videos or a playlist into structured knowledge. It can process videos at different detail levels and optionally export entities as Obsidian notes.

In plain words
What is it for?
Use it to process individual videos or playlists, choose transcription depth and engine, and create Obsidian notes from extracted entities.
Why use it?
It removes the need to handle each video manually and gives repeated processing a defined workflow.

Command for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/velmighty/youtube-to-knowledge/process
Clone the repo
git clone --depth 1 https://github.com/velmighty/youtube-to-knowledge

Made for: Claude Code.

Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,303 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.01303
Opus 5 $0.00000 $0.00651
Sonnet 5 $0.00000 $0.00261
Haiku 4.5 $0.00000 $0.00130

Measured 2d ago against content hash b695785c5697, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

process scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/commands/process.md · 132 lines

How it starts

The opening of the file, as written. The whole thing — 132 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Process one or more YouTube videos (or a playlist) into structured knowledge.

Arguments: $ARGUMENTS

Parse $ARGUMENTS to extract:

  • URLs: All arguments that start with http or look like YouTube video IDs. There may be one or many.
  • Playlist: If a URL contains playlist?list=, it is a playlist — expand it first (see Step 0).
  • --depth: light | standard | deep (default: standard).
  • --engine: whisperx | whisper (default: whisper).
  • --obsidian: flag (no value). If present, export entities as Obsidian markdown notes.

Examples:

  • /process https://youtube.com/watch?v=xyz
  • /process https://youtube.com/watch?v=abc https://youtube.com/watch?v=def
  • /process https://www.youtube.com/playlist?list=PLxxx
  • /process --depth deep https://youtube.com/watch?v=xyz
  • /process --obsidian https://youtube.com/watch?v=xyz

Follow these steps exactly:

Step 0 — Expand playlist (only if a playlist URL was given)

Run: python src/playlist_extractor.py <playlist_url>

Each line of stdout is a video URL. Replace the playlist URL in your working list with these individual URLs. Proceed to Step 1 for each video sequentially.

Step 1 — Duplicate check

If processing multiple videos: announce "Processing X videos" before starting.

Read vault/processed_videos.md. Extract the video ID from the URL. If the ID appears in the file, skip this video (print "SKIPPED: already processed") and move to the next one.

Step 2 — Transcription

Run: python src/transcribe.py <URL>

Parse stdout for:

  • CHANNEL_DIR:<path>
  • RAW_FILE:<path>
  • SOURCE_LANG:<lang>
  • TITLE:<title>
  • ENRICHED_FILE:<path> (optional, only from WhisperX)

If the script exits with error code 1 (no transcript found):

  • If engine is whisper (default): run python src/transcribe_whisper.py <URL>
  • If engine is whisperx: run python src/transcribe_whisperx.py <URL>
  • If the chosen engine also fails, try the other as a last resort.

Parse the same output variables from whichever script succeeds.

Read the full file on GitHub · 132 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 132 lines · 0 tokens per session scan A b695785c5697

Subscribe to this mod's changes

process is a command published in the GitHub repository velmighty/youtube-to-knowledge (59 stars, last pushed 4mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,303 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.