agentleFS
sgaabdu4sgaabdu4/building-flutter-apps/.agents/skills/marketing-video/SKILL.md

marketing-video

Skill25 starsChanged 3 months agoMITView on GitHub

Make a voiced product marketing or explainer video from real product screens. Script, cut walkthrough clips or stage app screenshots with a cursor, pick voice takes, render branded motion graphics and diagrams, mix music and review every second before delivery.

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    All hold on the exact delivered file:

    line 51
  2. Distinctive:

    every audit sheet opened, flat runs explained, stills checked at clicks, zooms + graphic scenes.

    line 55

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Marketing Video

Output = real product screens + big brand type + one calm voice → reviewed MP4. Scripts run by path from any directory (bun <skill>/scripts/<name>.mjs <work>); each video's work folder lives outside the skill.

Routes

Load only the routes the task reaches.

flowchart LR
  T{Task} -->|Runtime / commands / work folder| S[setup]
  T -->|New video| B[brief + script]
  B -->|Recordings exist| RC[recordings]
  B -->|Private data / no recordings| ST[staged screenshots]
  RC --> V[voice]
  ST --> V
  V --> R[render, review, deliver]
  T -->|Feedback on a draft| E[revisions]
  E --> R
  click S "references/setup.md"
  click B "references/script.md"
  click RC "references/recordings.md"
  click ST "references/staged.md"
  click V "references/voice.md"
  click R "references/render.md"
  click E "references/revisions.md"

Recordings come from Product Walkthrough Video takes or existing product videos.

Core rules

  • Agent cannot hear → voice, pace + pronunciation = human's call. One sample before every line is generated; never call audio "better" without their listen.
  • Real product only: screens, logos + fonts from the product and brand owners. Screens keep their own colours; the brand palette dresses only the graphics around them. Logo file, never a typed brand name; never an invented screen or feature.
  • Cut sign-in, invite, loading + empty screens. Access behaviour only when the audience must learn it (training); never a typed password.
  • Private data never appears: demo accounts or made-up records.
  • Zoom = hand-placed box on what the voice names, whole box in frame. No pointer-following or automatic zoom; a staged cursor never drives the camera.
  • Recorded clips → hard cuts. Staged shots/scenes → 0.4 s crossfades allowed. No slow fades to or from black.
  • Explanatory graphics copy an approved brand design (deck slide, site section) or the template defaults; no new card, chip or panel styles.
  • Plain, specific words, like a colleague showing the product. No hype, slogans or rule-of-three taglines; boldness = type size, layout, motion.
  • Human-approved part → frozen. Change only what the latest feedback names.

Completion gate

All hold on the exact delivered file:

  • script approved by the human; every claim backed by product material;
  • build.mjs passes; render.mjs reported no page errors;
  • every audit sheet opened, flat runs explained, stills checked at clicks, zooms + graphic scenes;
  • mix transcript holds every line in order;
  • delivered with draft number + change summary; voice quality stated as the human's judgment; music licence status stated.

Human step not done (no script approval, no voice listen) → deliver as a draft that names each unmet item; never call it complete.

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.