youtube-deepdive

youtube-deepdive is a command for Claude Code from hamzafarooq/claude-code-starter. It costs 0 tokens per session (466 once invoked), scanned A, original, MIT.

A command for researching one or more YouTube videos using their pages, descriptions, chapters, and comments. It produces a written guide to the video’s main sections and useful timestamps.

In plain words
What is it for?
Use it to summarize videos, find chapter times, identify useful tips in comments, and focus the review on a specific part or topic.
Why use it?
It saves time when you need the important parts of a long video without watching everything or searching through comments yourself.

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/hamzafarooq/claude-code-starter/youtube-deepdive
Clone the repo
git clone --depth 1 https://github.com/hamzafarooq/claude-code-starter

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for youtube-deepdive

README.md
[![agentmods](https://agentmods.dev/badge/commands/hamzafarooq/claude-code-starter/youtube-deepdive.svg)](https://agentmods.dev/commands/hamzafarooq/claude-code-starter/youtube-deepdive)
Your own site
<a href="https://agentmods.dev/commands/hamzafarooq/claude-code-starter/youtube-deepdive"><img src="https://agentmods.dev/badge/commands/hamzafarooq/claude-code-starter/youtube-deepdive.svg" alt="Measured on agentmods" height="20"></a>
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 466 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 $0.00466
Opus 5 $0.00000 $0.00233
Sonnet 5 $0.00000 $0.00093
Haiku 4.5 $0.00000 $0.00047

Measured 4d ago against content hash 72fb7ac60cb3, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

youtube-deepdive 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 4d 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/youtube-deepdive.md · 49 lines

What it actually says

If one or more YouTube video URLs were not provided, ask the user:

  1. Which YouTube video(s) they want to deep dive (paste the URL(s))
  2. Whether they want to focus on a specific aspect (e.g. "just the tips", "only the first half")

Then for each video URL:

  1. Use Brave MCP to navigate to the video page — mcp__brave-devtools__browser_navigate
  2. Take a snapshot to read the page — mcp__brave-devtools__browser_snapshot
  3. Look for chapters/timestamps in the description (lines like 0:00 Intro, 3:45 Key concept)
  4. If the description is collapsed, click the expand button to reveal it
  5. Extract top comments that mention specific tips, moments, or timestamps
  6. Repeat for each video, then close the browser — mcp__brave-devtools__browser_close

Produce the following write-up directly — no preamble:

YouTube Deep Dive: [Video Title]

Channel: [channel name] | URL: [link] | Duration: [duration if visible]

What it's about

1-2 sentences from the description. Plain English.

Best parts to watch
Timestamp What happens Why it matters
0:00 ... ...
2:15 ... ...

(If no timestamps are found, summarize key sections based on the description.)

What viewers found valuable

2-3 highlights from top comments — brief quotes, noting what they're reacting to.

Should you watch it?

One sentence verdict — who this is for and whether it's worth the full runtime.


(Repeat the block above for each video provided.)


Rules:

  • Use plain English throughout — written for PMs, not engineers
  • Do not fabricate timestamps or comments — only report what you read from the page
  • If a video is private, deleted, or inaccessible, say so clearly and skip it
  • If no chapters exist, infer structure from the description text
  • After the write-up, ask: "Want me to save this to docs/youtube-deepdive.md?"

$ARGUMENTS

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. 4d ago First seen · 49 lines · 0 tokens per session scan A 72fb7ac60cb3

Subscribe to this mod's changes

youtube-deepdive is a command published in the GitHub repository hamzafarooq/claude-code-starter (139 stars, last pushed 1mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 466 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.