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.
git clone --depth 1 https://github.com/jonny-1812/demo-video-factoryWrote 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/commands/jonny-1812/demo-video-factory/demo)<a href="https://agentmods.dev/commands/jonny-1812/demo-video-factory/demo"><img src="https://agentmods.dev/badge/commands/jonny-1812/demo-video-factory/demo/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/commands/jonny-1812/demo-video-factory/demo"><img src="https://agentmods.dev/badge/commands/jonny-1812/demo-video-factory/demo.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 | $0.00022 | $0.03363 |
| Opus 5 | $0.00011 | $0.01682 |
| Sonnet 5 | $0.00004 | $0.00673 |
| Haiku 4.5 | $0.00002 | $0.00336 |
Grade A, and why
demo 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 11d 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 — 140 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Generate a custom ~36s SaaS demo video for: $ARGUMENTS
This runs entirely in THIS Claude Code session — no API key, no spawned claude -p. The video's visual quality comes from fixed, high-quality scene templates (src/templates/scenes.tsx). Your only job is to write a brief.json (DATA) — never write scene code. Do everything yourself; don't delegate.
The brief feeds four templated scenes: PAIN → REVEAL → WOW → OUTCOME. The single biggest quality lever is using the product's real screenshots as scene heroes.
Step 1 — Scan + scrape (Bash)
npx tsx agent/scan.ts $ARGUMENTS # Playwright: real screenshots + brand colors/fonts
npx tsx agent/scrape.ts $ARGUMENTS # text content
Note the slug. Read public/real/<slug>/manifest.json — note brand (real colors/fonts), heroCandidates (real screenshot paths, screenshots first), and textOnly.
Use the Read tool to LOOK at the hero candidates (public/real/<slug>/hero.png, section_0.png, img_0.png, …). Pick the ONE screenshot that best shows the actual product UI — that's your hero for scenes 2/3/4. Also read out/<slug>_scraped.json for copy.
If manifest.textOnly is true (site blocked the scanner, e.g. Ramp): tell the user the site served a text-only page so there are no real product visuals, and the demo will be weak unless they drop real screenshots into public/real/<slug>/. Proceed only if they want a no-real-image version.
Step 2 — Write the brief (YOU write out/<slug>_brief.json)
Pick brand colors from manifest.brand (real site colors). isDark = is the site dark? (The product's real logo is auto-captured during the scan — manifest.logo — and shown on the Reveal + Outcome scenes automatically; you don't need to add it to the brief. If manifest.logo is null the scenes fall back to a text wordmark.) Pick fonts: one of inter | playfair | grotesk | dmsans (playfair for warm/editorial/lifestyle, grotesk for techy, inter for most, dmsans for friendly). Language/RTL: if the scanned site is in Hebrew (or another RTL/non‑Latin script the Latin fonts can't render), set fonts.heading and fonts.body to heebo (Hebrew + Latin) and write the on‑screen copy in that language — a native‑language demo is far stronger than anglicizing it. Keep the company name as‑is. Write product-specific copy (real, specific — no "saves time", no placeholders). Set every hero to your chosen real screenshot path (e.g. real/<slug>/hero.png).
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.
- 11d ago First seen · 140 lines · 22 tokens per session scan A fede70a5568d
demo is a command published in the GitHub repository jonny-1812/demo-video-factory (1 stars, last pushed 2mo ago), licensed MIT. It adds 22 tokens to every session and 3,363 once invoked, about $0.0001 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 commands, from other repositories
rotate-creative
Pause fatigued ads and ship fresh creative variants in their place. Works on both Meta and Google Ads. Use when CTR is dropping, frequency is climbing, or the user says "creative is tired" / "let's refresh ads". Always starts new variants paused.
review
Cold re-quiz on code that already shipped — your own session commits, not the change in front of you.
status
Show the gate's current state — armed or not, difficulty, marker, streak, open concepts.
setup
Set up PromptCut - installs deps (pip, ffmpeg), checks keys, CapCut folder.
video
Make a full video from an idea or a script - storyboard, voiceover, images, music, subtitles, render.
shorts
Cut vertical shorts or reels - from a topic, a script, or an existing long video.