antvis/chart-visualization-skills
7 skills8.6k installs2.4k starsGitHub
Install
npx skills add https://github.com/antvis/chart-visualization-skillsSkills in this repo
1Chart VisualizationChart Visualization is an agent skill that maps your data and intent to an AntV chart type, builds the JSON payload, and calls the gpt-vis API so you get a hosted chart image without hand-coding D3 or chart components. Solo builders use it when they need quick bar, line, pie, scatter, sankey, mind-map, or flow visuals for pitch decks, README metrics, internal dashboards, or growth snapshots. The skill enforces a short pipeline: analyze data, choose the chart from its decision guide (e.g., line for trends, bar for categories, pie for composition), curl POST to antv-studio.alipay.com with source set correctly, then paste the Markdown image link. That keeps agents from improvising unsupported chart types or malformed bodies. It matters because clear charts speed validation and communication when you are one person presenting numbers to users or investors. Pair it with your own data cleaning; the skill does not replace a full embedded chart library in production if you need interactivity.3.1kinstalls2Infographic Creatorinfographic-creator is an agent skill that turns plain-language briefs into AntV Infographic configurations solo founders can render or hand to a pipeline. Instead of guessing chart types ad hoc, the agent follows the Infographic DSL—indentation-based template, data, and theme sections aligned with AntV Infographic semantics. That matters for indie operators who need one visual to explain pricing tiers, feature comparisons, or research takeaways without hiring an illustrator. The skill front-loads specification reading: template names express information architecture, data carries title, description, and list rows with labels and values, and theme controls palette and typography. It is generative media work, not a live analytics dashboard; output is configuration text ready for the AntV renderer. Use when someone says they want an infographic from existing notes, not when they need interactive BI embeds.1.4kinstalls3Narrative Text VisualizationNarrative Text Visualization is an agent skill that walks solo builders through producing insight-driven stories from data using T8 Syntax—AntV’s declarative language where text carries structured semantics and embedded mini visualizations. Instead of hand-building DOM or chart configs for every paragraph, you describe the narrative and entities you want surfaced, then render with standardized styling across frameworks. Use it when users ask for interpretation reports, executive summaries, or structured articles that explain metrics in prose plus charts. The workflow starts from requirements gathering through T8 generation, making it LLM-friendly for agent-assisted analytics content. It suits product analytics writeups, research digests, and dashboard-adjacent explainers during Build, and reused narrative templates during Grow content or Launch GEO-style explainers.978installs4Icon RetrievalIcon retrieval is a small agent skill that wires your coding agent to an HTTP icon search API and standard curl fetches so you can drop SVGs into charts, landing sections, or AntV-style infographics without leaving the repo. Solo builders use it when a layout needs consistent pictograms—security, documents, data—and you want the agent to resolve keywords to retrievable SVG strings instead of guessing filenames or bundling a full icon font. The flow is deliberate: pick a concept keyword, call the search endpoint with text and topK, parse the JSON payload for URL list, then fetch each SVG with curl for paste-in or component embedding. It fits the build phase for anyone shipping visual product surfaces who already trusts shell and network access in the agent. Complexity stays beginner-friendly because the API surface is one GET and one asset fetch pattern; confidence is high where the endpoint is reachable. It is not a general image CDN or licensing catalog—only SVG-oriented icon search through the documented host.934installs5Antv S2 ExpertAntV S2 Expert is an agent skill for solo builders embedding Ant Financial’s S2 pivot and table engine in TypeScript apps. It focuses on advanced customization—subclassing DataCell (or TableDataCell for flat tables), hooking custom draw methods, and registering cells through S2Options so agents do not guess AntV G APIs from scratch. Typical jobs include finance ops dashboards, SKU matrices, cohort pivots, and admin reporting where default cells are insufficient and you need images, badges, or bespoke backgrounds inside grid cells. The skill aligns with Build-phase frontend work alongside @antv/s2 and @antv/g, and suits Claude Code or Cursor users shipping internal analytics for a SaaS or API product. It is integration-oriented rather than a full chart catalog; pair it with broader AntV chart skills when you also need line or bar specs outside the pivot model.811installs6Antv Skills MaintainerAntV Skills Maintainer is an internal automation skill for the AntV chart-visualization skills repository. It is meant for maintainers who add or edit user-facing skills and need README “Available Skills” and `.claude-plugin/marketplace.json` to stay accurate without manual copy-paste. The skill documents a fixed sync ritual: emoji + bold name, one-line description from frontmatter, and a short capability paragraph per skill, while keeping the maintainer skill itself out of public indexes. Solo builders browsing Prism should treat this as maintainer tooling, not a charting or visualization capability. If you fork a similar skills monorepo, the pattern is reusable; the AntV-specific paths and “do not list maintainer” rule are the critical constraints. Intermediate complexity because it assumes familiarity with plugin marketplaces and repo layout.725installs7Antv G2 Chartantv-g2-chart is procedural guidance for AntV G2 version 5’s animation system, aimed at solo builders who ship analytics or admin dashboards and want motion that reads clearly instead of distracting. You configure the `animate` field for three lifecycle moments—enter when data first paints, update when values change, and exit when series or marks leave the view. The skill catalogs built-in presets such as fade, axis-aligned scale and grow, wave sweeps for polar charts, zoom for scatter points, path drawing for lines, and morphing when chart types change. Each preset can be tuned with duration, delay, and easing so transitions stay consistent with your product’s feel. It pairs with broader G2 chart-init workflows and assumes you already have a container and dataset; use it when first render or data refresh looks abrupt, or when you need a quick reference for which preset fits bars versus pies versus polylines.586installs