A command for analyzing a video from a YouTube link, local file, or directory. It produces a summary, key points, topics, sentiment, important timestamps, and visual markers for selected moments.

Use it to analyze one video or a directory of videos, capture notable screen content and slides, and record where important moments occur.

It turns a video into searchable notes so you can understand its content without manually reviewing the entire recording.

video — Command for Claude Code by Galbaz1/video-research-mcp

Install

Getting it into your agent

It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.

Clone the repo

git clone --depth 1 https://github.com/Galbaz1/video-research-mcp

agentmods

npx agentmods add commands/galbaz1/video-research-mcp/video

Made for: Claude Code.

Or install gr, the plugin that ships this one along with the rest of its 27 skills, 17 commands, 7 agents.

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.

README.md

[![agentmods](https://agentmods.dev/badge/commands/galbaz1/video-research-mcp/video/github.svg)](https://agentmods.dev/commands/galbaz1/video-research-mcp/video)

Your own site

<a href="https://agentmods.dev/commands/galbaz1/video-research-mcp/video"><img src="https://agentmods.dev/badge/commands/galbaz1/video-research-mcp/video/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/commands/galbaz1/video-research-mcp/video"><img src="https://agentmods.dev/badge/commands/galbaz1/video-research-mcp/video.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 14 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 2,287 The whole file, excluding the scripts and references it only reads on demand.

Security scan B 2 findings. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00014	$0.02287
Opus 5.5	$0.00006	$0.00915
Sonnet 5.5	$0.00003	$0.00457
Haiku 4.5	$0.00001	$0.00229

Measured 1mo ago against content hash 296ad468446f, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade B, and why

video scanned grade B with 2 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 1mo 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.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

a. Determine the memory directory: find the `.claude/` project memory path for the current working directory. Use `Glob` on `~/.claude/projects/*/memory/` to find the active project memory path if needed.

Runs shell commandslowCapability

Expected in a hook, worth knowing in a rule or an instructions file.

result = subprocess.run(['ffmpeg', '-y', '-ss', ffmpeg_ts, '-i', video_path, '-frames:v', '1', '-q:v', '2', out], capture_output=True)

commands/video.md · 223 lines

How it starts

The opening of the file, as written. The whole thing — 223 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Video Analysis: $ARGUMENTS

Analyze the provided video source with progressive memory saving and automatic visualization.

Phase 1: Analyze

Determine the input type from "$ARGUMENTS":
- If it starts with http:// or https://: use video_analyze with url parameter
- If it's a directory path (ends with / or is a known directory): use video_batch_analyze with directory parameter
- Otherwise (file path): use video_analyze with file_path parameter

For single video analysis, use this instruction (append the screenshot marker block for local files only):

Base instruction (always): instruction="Provide a comprehensive analysis including title, summary, key points, timestamps of important moments, main topics, and overall sentiment."

Screenshot markers (append when input is a local file, NOT a YouTube URL):

Also place inline screenshot markers at visually important moments using this format:
[SCREENSHOT:MM:SS:brief description of what is shown on screen]

Rules:
- Place 8-15 markers spread across the video
- Capture: screen shares, diagrams, slides, demos, key visual moments
- Skip: webcam-only talking heads, transitions, blank screens
- Description should explain what is visible (e.g., "architecture diagram showing data flow")

For batch (directory) analysis, use instruction="Provide a comprehensive analysis of this video."

Phase 1.5: Download to Memory (YouTube only)

If the source is a YouTube URL, persist a local copy for reuse.

Compute video_id from the URL and set target path: <memory-dir>/gr/media/videos/<video_id>.mp4

If the target file is missing, download with:

yt-dlp --no-playlist -q -f "mp4[height<=720]/mp4/best[ext=mp4]" \
  -o "<memory-dir>/gr/media/videos/<video_id>.mp4" \
  "https://youtube.com/watch?v=<video_id>"

Update <memory-dir>/gr/media/videos/.manifest.json with title, source_url, size_mb, downloaded_at.
If download fails or yt-dlp is unavailable, continue analysis without local media.

Read the full file on GitHub · 223 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.

1mo ago Changed · +5 lines 296ad468446f
1mo ago Changed · -1 lines 1cc3e159423f
2mo ago First seen · 219 lines · 14 tokens per session scan B 8339e08737c6

Subscribe to this mod's changes

video is a command published in the GitHub repository Galbaz1/video-research-mcp (23 stars, last pushed 6d ago), licensed MIT. It adds 14 tokens to every session and 2,287 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it B with 2 findings (reads agent configuration directories, runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Claude Code Creative and media Research ai-tools gemini mcp mcp-server

video next to 6 commands from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
videoGalbaz1/video-research-mcp	Command	23	6d ago	MIT	B	14	Claude Code
setup-video-visionjordanrendric/claude-video-vision	Command	1.5k	23d ago	MIT	A	20	Claude Code
eg-new-featurevshvedov/elephant-goldfish	Command	92	2d ago	MIT	A	20	Claude Code
eg-fix-bugvshvedov/elephant-goldfish	Command	92	2d ago	MIT	A	23	Claude Code
dock-chatJacksunwei/claude-telegram-buddy	Command	3	5mo ago	Apache-2.0	A	12	Claude Code
mem-lastdjinn-soul/gemini-mem	Command	3	6mo ago	MIT	A	0	Claude Code
mem-prunedjinn-soul/gemini-mem	Command	3	6mo ago	MIT	A	0	Claude Code

video