compress-video · pixel-point — Claude Code skill for video compression
The compress-video Claude Code skill compresses videos and compares codec and CRF quality. It can also inspect embedded website videos and optimize assets across a source project.
MIT ★ 64 updated 2026-11-06
npx skills add pixel-point/skills --skill compress-video -a claude-codeWhat compress-video does
The skill compresses local files with FFmpeg and FFprobe, and can estimate quality and output sizes for a comparison. A request to process a file starts the compression workflow unless a different action is requested.
For a website audit, it examines source code and embedded videos on a page or within a specified scope. Browser tools and source-code access are needed to check the rendered size and playback. When no URL is supplied, it reuses a development server or starts the project locally.
The optimization workflow scans the whole project by default, unless the request limits its scope. FFmpeg and FFprobe are required for video work; the skill installs dependencies needed for the selected workflow when they are missing. A site audit cannot begin without access to its source code.
What it helps with
- compress a local video file
- compare video quality and estimated output sizes
- inspect embedded videos on a website page
- scan a project and optimize its video assets
Example prompts for the agent
- Compress
demo.mov. - Compare quality and estimated size for
demo.mov. - Scan
./websiteand optimize its videos usinghttp://localhost:3000.
How to install compress-video
1Skills CLI (Vercel)
npx skills add pixel-point/skills --skill compress-video -a claude-code2Manually: put the files in place
~/.claude/skills/compress-video/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/compress-video/
1Skills CLI (Vercel)
npx skills add pixel-point/skills --skill compress-video -a codex2Manually: put the files in place
.agents/skills/compress-video/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add pixel-point/skills --skill compress-video -a cursor2Manually: put the files in place
.cursor/skills/compress-video/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install pixel-point/skills skills/compress-video2Manually: put the files in place
.github/skills/compress-video/SKILL.md1Skills CLI (Vercel)
npx skills add pixel-point/skills --skill compress-video -a gemini-cli2Platform CLI
gemini skills install https://github.com/pixel-point/skills --path skills/compress-videoManually: put the files in place
.devin/skills/compress-video/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about compress-video
How do I install compress-video in Claude Code?
Run npx skills add pixel-point/skills --skill compress-video -a claude-code. The skill file is at ~/.claude/skills/compress-video/SKILL.md.
What does compress-video do?
It compresses local videos, compares quality and estimated sizes, and can inspect embedded website videos or optimize video assets across a source project.
What does compress-video need to run?
Video workflows use FFmpeg and FFprobe; the skill installs missing dependencies needed for the requested workflow. Website audits also need source-code access and browser tools.
Source text
SKILL.md · MIT
--- name: compress-video description: Compress local videos with FFmpeg, compare codec/CRF quality, audit embedded videos on a website, or scan and optimize videos across a website's source project. --- Compress videos locally using FFmpeg and FFprobe. Install missing dependencies needed for the requested workflow. Website workflows require source-code access and browser tools to verify the rendered site. ## Commands | Request | Workflow | | --- | --- | | A video file, or `compress <file>` | Compress using [compression.md](references/compression.md). | | `compare <file>` | Compare quality and estimated sizes using [comparison.md](references/comparison.md). | | `audit [project] [URL or scope]` | Inspect embedded videos and report improvements using [audit.md](references/audit.md). | | `scan-and-optimize [project] [URL or scope]` | Discover, audit, optimize, and update videos across the project using [scan-and-optimize.md](references/scan-and-optimize.md). | Infer the workflow from natural-language requests too. Read only the relevant references. A file without a command uses the compression workflow. For both website workflows, use the current workspace or a supplied project path. If source code is unavailable, ask for repository access before starting. Begin with routes, components, styles, and asset references in the source code; use a live or local URL to verify rendered video sizes and playback. If no URL is supplied, reuse a development server or start the project locally. `audit` reviews the requested page and its source code, or the specified scope. `scan-and-optimize` covers the whole source project by default; a supplied URL is its browser starting point. Honor an explicitly narrower scope. Ask for a missing input only when it cannot be inferred. Reuse source locations and preferences already provided. Compression defaults, including CRF values, are defined in [compression.md](references/compression.md). ## Examples ```text $compress-video demo.mov $compress-video compare demo.mov $compress-video audit ./website https://example.com/blog/post $compress-video scan-and-optimize ./website http://localhost:3000 ``` These are agent requests, not terminal commands.
© pixel-point · MIT · pixel-point/skills
Similar skills
All skills for Agent setup: memory, context, orchestration →compress-video by other authors comes first.