agentleFS
frankxaifrankxai/agentic-creator-os/.claude/skills/apod-to-short/SKILL.md

apod-to-short

Skill10 starsChanged 7 months agoApache-2.0View on GitHub

Turn a NASA Astronomy Picture of the Day (APOD) into a 30–60s vertical short — script, shot list, captions, and a rights line. Use when the user mentions APOD, 'astronomy picture of the day', or wants a daily space short/reel/TikTok from a NASA image.

Files
SKILL.md

APOD → Short

NASA Astronomy Picture of the Day → a ready-to-shoot 30–60s vertical short.

Purpose

NASA publishes a stunning image with an expert explanation every day. This skill converts one APOD entry into a complete vertical short package: a hook, a tight narration script, an on-screen caption track, a shot/animation list, and a rights line — so a creator can film/edit in one sitting without misstating the science or the licensing.

When it fires

  • Keyword triggers: apod, astronomy picture of the day, daily space short, nasa image short, space reel
  • File triggers: apod.json, apod-*.md
  • Command triggers: /apod-to-short

Inputs

InputRequiredNotes
apodyesAPOD URL/date, or the title + explanation text + image credit.
lengthnoTarget seconds, default 45 (range 30–60).
voicenostarlight (default), arcanea, or frankx — see brand skills.

Workflow

  1. Resolve the APOD — get title, date, explanation, media URL, and the exact image credit. If credit names a non-NASA party (ESA, an observatory, an individual), flag it for the rights step. ✅ Check: you have a verbatim credit string.
  2. Extract the one idea — reduce the explanation to a single, true, surprising fact a non-expert can grasp. ✅ Check: it survives "what does that mean specifically?"
  3. Write the hook (0–3s) — a question or claim that earns the first three seconds without clickbaiting the science.
  4. Script the body — narration timed to length at ~2.5 words/sec. Keep every claim traceable to the explanation; never invent figures.
  5. Build the caption track — short on-screen lines synced to beats (≤7 words each).
  6. Write the shot list — Ken Burns push/pan on the still, plus any overlay labels (object name, distance, instrument). No fabricated footage.
  7. Emit the rights line — exact attribution; route through rights-check-nasa-esa if credit is non-NASA.

Output contract

{
  "title": "string — APOD title",
  "hook": "string — 0–3s opener",
  "script": "string — full narration, timed",
  "captions": ["string — ≤7 words each"],
  "shot_list": ["string — shot/animation directions"],
  "rights_line": "string — exact attribution",
  "duration_sec": "number"
}

Tools & MCP

  • Tools: WebFetch (resolve APOD), Read.
  • MCP dependencies: none required. (Optional: a NASA-media MCP from starlight-mcp to fetch APOD JSON directly.)

Quality bar

  • Every stated fact is traceable to the APOD explanation — no invented numbers.
  • Rights line is verbatim and correct; non-NASA credit is flagged.
  • Narration fits length at ~2.5 words/sec.
  • Hook respects the audience's intelligence (no "scientists HATE this").

Example

See examples/example-01.md.


Built on SIP — Starlight Intelligence Protocol Substrate: starlightintelligence.org/protocol v1.1.0 Layers used: [file-contract, attestation, sovereignty] Vertical: starlight-agent-skills · portable capability layer

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.