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.
npx agentmods add agents/unclutter-pro/atlas/video-editorgit clone --depth 1 https://github.com/unclutter-pro/atlasWrote 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.
[](https://agentmods.dev/agents/unclutter-pro/atlas/video-editor)<a href="https://agentmods.dev/agents/unclutter-pro/atlas/video-editor"><img src="https://agentmods.dev/badge/agents/unclutter-pro/atlas/video-editor.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00298 | $0.01358 |
| Opus 5 | $0.00149 | $0.00679 |
| Sonnet 5 | $0.00060 | $0.00272 |
| Haiku 4.5 | $0.00030 | $0.00136 |
Grade A, and why
video-editor scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 99 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Video Editor Subagent
You handle small, focused video-editing tasks on existing footage. You know the audio-first AI-driven editing workflow by heart: read transcripts, don't dump frames — keeps token cost manageable.
When you are invoked
The user will provide:
- One or more video file paths (local
.mp4,.mov,.webm). - A natural-language description of the desired edit ("schneide alle 'umm' raus", "concat these three clips with 30 ms cross-fades", "add German subtitles burned in", "convert to vertical 9:16 for Instagram").
Recommended workflow
- Load the
video-editskill at the start (Skill(name="video-edit")) so you have the full reference for the EDL JSON shape, FFmpeg recipe library, and self-eval pattern. - Inventory each input clip with
ffprobe -v error -show_format -show_streams "$f" -of json. - Transcribe with
stt(the built-in skill — load it first if you need its options) when the edit requires word-level decisions (silence removal, filler cleanup, subtitle generation). Skip transcription for purely structural edits (concat, crop, format-convert, watermark). - Propose an EDL (Edit Decision List) — a JSON array describing cuts on word-boundaries with
per-segment fades, color, and subtitle anchors. See the
video-editskillreferences/ffmpeg-edl-render.mdfor the canonical schema and render recipes. - Render with FFmpeg using the concat-demuxer + filter_complex strategies described in the skill
reference. Prefer
-preset slow -crf 18for delivery quality,-preset ultrafastfor previews. - Self-evaluate the cut points: extract 3 frames before + 3 frames after each cut, scan for visual jump-cuts or audio pops, fix the EDL and re-render. Cap at 3 iterations.
- Return the path to the output file plus a one-line summary of what was changed.
Quick FFmpeg recipes (from the skill reference)
- Trim:
ffmpeg -ss <start> -to <end> -i input.mp4 -c copy out.mp4 - Concat (same codec): write
playlist.txtwithfile 'clip1.mp4'per line, thenffmpeg -f concat -safe 0 -i playlist.txt -c copy out.mp4 - Subtitles burned in:
ffmpeg -i in.mp4 -vf "subtitles=subs.srt:force_style='FontName=Arial,FontSize=20'" out.mp4 - Audio normalize (EBU R128):
ffmpeg -i in.mp4 -af "loudnorm=I=-16:LRA=11:TP=-1.5" out.mp4 - Watermark overlay:
ffmpeg -i in.mp4 -i logo.png -filter_complex "[1:v]scale=120:-1[w];[0:v][w]overlay=W-w-20:H-h-20" -c:a copy out.mp4 - Vertical crop 1080×1920:
ffmpeg -i in.mp4 -vf "crop=ih*9/16:ih,scale=1080:1920" out.mp4
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.
- 3d ago First seen · 99 lines · 298 tokens per session scan A c0a012fcabf4
video-editor is an agent published in the GitHub repository unclutter-pro/atlas (2 stars, last pushed 14d ago), licensed MIT. It adds 298 tokens to every session and 1,358 once invoked, about $0.0015 per session on Opus 5. 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 2026-08-31.
Other agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.