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 skills add andrewii23/ii23-skills --skill video-understandgit clone --depth 1 https://github.com/andrewii23/ii23-skillsWrote 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/skills/andrewii23/ii23-skills/video-understand)<a href="https://agentmods.dev/skills/andrewii23/ii23-skills/video-understand"><img src="https://agentmods.dev/badge/skills/andrewii23/ii23-skills/video-understand/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.
<a href="https://agentmods.dev/skills/andrewii23/ii23-skills/video-understand"><img src="https://agentmods.dev/badge/skills/andrewii23/ii23-skills/video-understand.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00191 | $0.02409 |
| Opus 5 | $0.00096 | $0.01205 |
| Sonnet 5 | $0.00038 | $0.00482 |
| Haiku 4.5 | $0.00019 | $0.00241 |
Grade C, and why
video-understand scanned grade C with 1 finding 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 12d 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
Clean up the work dir with `rm -rf` when the user is unlikely to follow up. How it starts
The opening of the file, as written. The whole thing — 177 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Video Understand
Read a video the way a person does: see the picture, hear the words, tie them to the clock. Two scripts produce that; you do the understanding.
scripts/frames.py— extract the frames that matter, drop the duplicates, pack them into grid images + a timestamp manifest.scripts/transcript.py— a timestamped transcript (ElevenLabs by default).
Both halves are required for a real answer. A grid shows what is on screen and never what was said. On a 24-minute episode the grid alone gives the arc (someone is accused, a gadget appears, a shadow chases them) while every fact that matters — what was stolen, the gadget's name and rule, how it resolves — lives entirely in the audio. Skipping the transcript produces confident, plausible, wrong summaries. Skip it only when the video has no speech, or the user explicitly asks for visuals only.
Resolve SKILL_DIR first
Every command below runs a script under SKILL_DIR/scripts/. Set SKILL_DIR to
the absolute path of the directory holding THIS SKILL.md (your harness reported
it when you read this file). Scripts are always a direct sibling of this file.
On Windows substitute python for python3.
Step 1 — frames
python3 "$SKILL_DIR/scripts/frames.py" "<video>" --out-dir <work-dir> --cells 64
Prints MANIFEST: <path>. Pick --cells by the QUESTION, not the video length —
the real limit is how many pixels each cell survives at after downscaling:
--cells |
px per cell | Use for |
|---|---|---|
| 16 | ~392 | small on-screen text: UI labels, timestamps, subtitles |
| 36 | ~261 | facial expression, fine action, motion beats |
| 64 | ~196 | plot, scene inventory, recaps, spoilers (default) |
| 100 | ~157 | rough skim of very long footage; small text is lost |
Text legibility, measured on 720p: big burned-in captions and full-screen title
cards (including dense Japanese) read fine at 196 px/cell. Player-UI text —
0:01 / 20:56, a chapter label — needs 392 px/cell. So reach for --cells 16 only for genuinely small text; 64 already handles anything title-sized.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago First seen · 177 lines · 191 tokens per session scan C c012949459a6
video-understand is a skill published in the GitHub repository andrewii23/ii23-skills (10 stars, last pushed 1mo ago), licensed MIT. It adds 191 tokens to every session and 2,409 once invoked, about $0.0010 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…