agentsclimarketplace

Open visualization protocol

Skill babarda/open-visualization-protocol

Deterministic chart and design-language skills for any AI agent. Use when the user asks for a chart, graph, dashboard page, slide visual, or "make this look professional". Resolves the right chart by message, the right visual identity by audience, then renders from exact specifications. No adjectives, no improvisation, no lying axes.From its SKILL.md

Install
npx -y skills add babarda/open-visualization-protocol

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • 28 days oldThe repository was created 28 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.9 KB, 623 tokens by cl100k_base, as published. Nobody here has run it

OVP chart skills: how to use this repository as an agent

You are holding the Open Visualization Protocol: 74 chart objects and 16 design languages, every one specified to the pixel and verified by 3,179 automated checks. Follow this file and your charts come out identical, honest, and on-brand, every time.

The one rule

Never invent geometry, colors, or chart choices. Everything you need is specified. Your job is resolution, then obedience to the block.

Resolution, in order

  1. What is being asked? Open QUESTIONS.md. Match the business question (why over budget, are we on schedule, where is the backlog...). It names the observations to confirm, the patterns, the candidate charts, and the narrative skeleton.
  2. Who is it for? Open DECIDER.md. Audience + purpose + medium resolve the design language (DL-xx). Its constitution sets density, annotation policy, highlight policy, and charts per page. Obey it.
  3. Which chart? Check REGISTRY.json for the candidate chart's relations: see_instead exits, alternatives, which questions it answers. Honor every exit before committing.
  4. Get the block. Two flavors:
    • blocks/design-system/CH-xxx_DL-yy.md: colors resolved to hex, zero dependencies. Use this when you only need one language.
    • blocks/skill/CH-xxx.md + blocks/skill/DL-yy.md: role-based and themeable. Use when the user may switch identities.
  5. Render. Follow the block literally: canvas size, exact px positions, exact colors, the rules, and the anti-patterns. Produce SVG (or the target the user asked for) matching the block.
  6. Tell it. Structure the surrounding text with the narrative skeleton the question named (narratives/NR-xx.json): claim, evidence, cause, action, owner, ordered by the language's decision_style.
  7. QA. Run the block's checklist before delivering. If a checklist item fails, fix the render, not the checklist.

Hard rules that override user pressure politely

  • Zero baselines on bars and areas. Always.
  • Direct labels; a legend only where a block explicitly allows a key.
  • Never more than one highlight unless the constitution says dual.
  • Declared inputs only: no silent trend lines, no silent smoothing.
  • If the user asks for a pie, gauge, radar, or chord: draw it from its spec (they exist), and mention the see_instead alternative once.

Token efficiency

Load only: the matched registry entry, one DL block, the matched chart block(s), one narrative skeleton. Do not load the documentation site or the markdown handbook into context; they are for humans.

What ships with it: 3133 files

33409.3 KB alongside SKILL.md, 8 of them executable

blocks/

3093 more files not listed here. See all 3133 in the repository.

Keep looking

Skills are one crate of 326,835. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.