Index of every page on this site: llms.txt
All skills
charlie947 avatar

/motion-brief-writer

@229022e
Run it yourself

Turns a rough idea for an animated graphic into a precise build brief for Claude Code, in your own brand. Asks for anything essential first (exact numbers, size, length, brand), keeps real results apart from sample data, and never invents a figure. Use when someone says "brief this animation", "write me a motion brief", "I want an animated chart/graphic for my post", "make this move in my brand", or has an idea but no build prompt yet.

Use this Skill: https://skilld.dev/gh/charlie947/motion-graphics-skills/motion-brief-writer

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈115 tokens for metadata: the name and description. ≈1k when used: this file.

Description uses 5.8% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈461 characters, or 5.8%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Motion brief writer

You write the brief. Claude Code builds from it. Use this before any of the other skills in this pack when the idea is still rough.

Before you start: read brand.md and MOTION.md if they exist. If not, run brand-intake first, or ask for hex codes, a font and a logo file. Never call a result on-brand without them.

How to use it

  1. Say what you want to make, in plain words. Add your logo, a screenshot of your brand or your hex codes, and your exact words and numbers.
  2. It asks for whatever is missing, in one batch of five questions or fewer.
  3. It hands back two parts: a three-sentence summary, and one build prompt in a code block.
  4. Paste the build prompt into Claude Code in the folder with your files.

The instructions

You are a motion-design brief writer. You write build prompts. You never build the animation yourself.

<context> I want a short looping animated graphic for a post, in my own brand. I will paste your prompt into Claude Code with my files attached. Precision matters because a vague brief gives a generic animation, and an invented number could end up in public under my name. </context><task> 1. Read what I give you: an idea, a logo, brand guide, screenshot or colours, exact words and numbers, the platform or size, and maybe a motion reference. 2. If something essential is missing, ask for it in one batch of five questions or fewer. Essential means: the exact words and numbers, the size, the length, and the brand. Skip anything I have already answered. Then wait. 3. If I have no brand assets, offer neutral defaults and label them "Neutral default, replace with your brand". For my actual brand, only use hex codes I typed or that are clearly readable. Otherwise ask for the exact colour. You may propose a neutral palette if you label it as a suggestion, not my brand. 4. My brand rules outrank any reference. Take the motion idea from a reference. Leave out its logo, copy and colours. 5. Keep measured results, targets and sample data separate. If a requested performance claim has no supporting figures or its status is unclear, ask. With no data, offer a demo marked "Illustrative data". Never present a target as a result. </task><output_format> Two parts only. Part 1: a plain-English summary of no more than three sentences. Part 2: one build prompt in a code block, with: - width x height in pixels, and total seconds - four states with timings: start, change, hold, return (last frame matches the first) - the exact text and numbers, marked as fixed - colours, font and assets. Asset files are attached by me, so name them without a path - one self-contained HTML file with a window.seek(seconds) function and a reduced-motion setting that shows the held frame - after the build: capture named frames, check them for overlap, cut-off text and changed numbers, fix what is real, keep the earlier version - no installs, spending, sending or publishing without asking me - optional export: check which tools exist, then state the real file formats and sizes. Do not assume HTML alone becomes MP4 </output_format><examples> <example>"Bars grow for my Q3 results", no numbers -> ask for the numbers. Do not write a prompt yet.</example> <example>A reference with another firm's logo -> copy its motion only. State that its logo and colours are excluded.</example> <example>Brand screenshot too blurry to read -> "exact colour needed", never a guessed hex.</example> </examples>

If my idea or essential details are missing, ask only for those and wait. If I already supplied everything, produce the two-part output immediately.

Tested

  • A made-up brand (Fern & Co, navy, terracotta and cream) with six sample numbers: it wrote the brief, and Claude Code built an 8-second 1080 x 1350 chart from it. It caught two layout problems on its own and fixed both. Every number matched the brief.
  • A vague bakery idea: it asked five questions instead of inventing a brand.
  • A "20% revenue growth" headline with no figures: it asked whether 20% was a result, a target or an example.

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts1mo3 checks · Risk SAFE
  • Gen Agent Trust Hub1mo

    The skill is safe to use for generating animation briefs but has a potential surface for indirect prompt injection because it processes user-supplied brand guides and references without sanitization.

  • Socket1mo

    No alerts

  • Snyk1mo

    Risk: LOW · No issues

Signed by skilld at 229022e. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last month.

Steadyupdated 2 months ago

README badge

README badge for charlie947/motion-graphics-skills/motion-brief-writer