What does this skill do, and when should you use it?
This is the official Agent Skill shipped inside TradingView's lightweight-charts repository, teaching AI assistants how to work with the small, fast HTML5-canvas financial charting library. It covers the v5 API conventions, an eight-layer mental model (chart, series, scales, data, interaction, panes, extensions, wrappers), and the frequent foot-guns around time handling, scales, markers, plugins, and framework wrappers. It insists on verifying every API name against local typings (node_modules or the upstream dist/typings.d.ts) instead of regenerating stale v4 snippets from memory. It ships copy-paste recipes for realtime updates, dual price scales, volume panes, timezone display, watermarks, and Next.js client-only components.
- Detects whether the project is an npm consumer or upstream checkout and resolves API names from local typings
- Maps v4-era snippets to their v5 replacements (addSeries, createSeriesMarkers, watermark APIs, etc.)
- Produces minimal runnable snippets for time scales, panes, per-bar colors, and coordinate conversion
- Adapts chart lifecycle and loading guidance to React/Vue/Web Components/Next.js SSR environments
- Routes plugin, custom-series, and rendering questions to official docs paths and plugin-examples/
- States what could not be verified locally and refuses to invent option names or methods
- A frontend developer embedding interactive candlestick/area charts with streaming realtime data
- Migrating v3/v4 code to v5 and needing confirmation of renamed APIs like addSeries and marker primitives
- Wrapping charts in React or Next.js (SSR) and handling useEffect lifecycle with client-only loading
- Building custom plugins or drawing tools that need primitive vs custom-series guidance and coordinate conversion
- Debugging timezone display, UTCTimestamp seconds-vs-milliseconds, or ascending-unique data requirements
- Users of the third-party Python lightweight_charts wrapper — the skill explicitly flags that as a separate package and defers core v5 JavaScript advice
- Server-side chart image rendering or non-browser scenarios; the library targets browser HTML5 canvas (iOS/Android wrapper docs exist, but the skill's focus is Web)
How do you install this skill?
- This is a static source review; no commands or code examples were executed and example correctness is unverified at runtime.
- No Chinese-language support is declared; the plain-HTML recipe and several referenced resources depend on unpkg, GitHub, and overseas sites — mainland-China reachability must be confirmed independently.
- Repo paths referenced in the routing table (website/docs, plugin-examples, etc.) were not individually verified in this evidence; consumer projects should rely on local node_modules typings first.
- Publisher is TradingView's official repository but unverified by the FollowSkills registry; the skill has no independent version or changelog — re-check currency against future v5.x releases.
- Shell / CLI
- Local filesystem
Node.js (npm-installed lightweight-charts package or upstream checkout)
Install into an Agent Skills-compatible assistant such as Claude Code or Codex:
npx skills add https://github.com/tradingview/lightweight-chartsHow do you use this skill?
Once installed, send your agent any of these to trigger it:
- Why does my chart snap to the right edge every time new data arrives? Fix it with incremental updates
- Migrate this v4 addCandlestickSeries and setMarkers code to lightweight-charts v5
- Create a client-only Next.js component for Lightweight Charts with auto-sizing
- How do I color individual histogram bars in a volume series?
Once installed, any conversation touching lightweight-charts chart creation, series, scales, markers, plugins, or wrappers triggers the skill's flow: it first runs verification commands (node -p to read the installed package version, rg against dist/typings.d.ts), then answers with a minimal v5 snippet, calls out the main foot-gun, and reports which local sources were checked.
What are this skill's strengths and limitations?
- Ships with the official TradingView repository, so it stays aligned with the library itself
- Forces local typings verification, sharply reducing invented APIs and outdated v4 snippets
- Includes a complete v4→v5 replacement matrix and an itemized foot-gun list
- Provides over a dozen copy-paste recipes (realtime updates, dual axes, watermarks, Next.js, etc.)
- Only covers lightweight-charts v5 conventions; useless for other charting libraries
- Full API verification requires access to node_modules or the upstream checkout; otherwise only partial verification is possible
- The third-party Python lightweight_charts wrapper is out of scope
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| TradingView Lightweight Charts Skill this page | 68 · Recommended | ★ 18k | 12d ago | Apache-2.0 |
| Effect JSDoc Authoring Skill | 77 · Strongly recommended | ★ 28k | 11d ago | MIT |
| FreeHire Tech Job Search Skill | 63 · Recommended | ★ 47k | 17d ago | MIT |
| Vibeyard UI Development Skill | 56 · Use with care | ★ 1.4k | 1mo ago | MIT |
| Codex CLI Builder ✓ OpenAI · Official | 55 · Use with care | ★ 28k | 4mo ago | — |
The README positions the library as among the smallest and fastest HTML5 financial charts, near the size of static images, and points to the awesome-tradingview community list for related projects; the skill itself does not compare against other Agent Skills.
How did FollowSkills review this skill?
Pure knowledge skill: no scripts, network calls, or side effects — inherently least privilege. It embeds verify-before-answer commands (local typings/package.), mandates honest disclosure when files are unavailable, forbids inventing APIs, and disambiguates the third-party Python wrapper, giving good data-flow transparency. Deductions: static review cannot confirm runtime behavior; no confirmation/rollback concepts (largely N/A for docs), and publisher is registry-unverified, so not full marks.
Internally consistent: version-detection order, v4→v5 replacement matrix, triage table, and clear failure guidance (state what could not be verified). Static cap of 10 applies; the repo has CI workflows (deploy.yml, pkg-pr-new.yml) and unit/e2e suites, but they cover the library, not the skill's key paths, so only a small uplift to 11.
Description and body match well: clear audience, trigger scenarios (series, scales, markers, plugins, wrappers), non-fit boundaries (Python wrapper; non-time horizontal axes require triage), and consumer/upstream environment detection. Deductions: no Chinese-language support note; reliance on unpkg/GitHub/official site without mainland-China reachability disclosure; no explicit false-trigger negative list beyond the wrapper check.
Well-layered docs: lookup order, mental model, triage, foot-guns, routing map, recipes. License (Apache-2.0) and version anchoring (v5, package. 5.2.1) are clear; upstream README provides install path (npx skills add) and maintenance ownership (TradingView). Deductions: no changelog/version history for the skill itself, no dedicated FAQ/known-limitations section, and many routing-table repo paths are not verifiable from the provided evidence.
Directly addresses high-frequency pain points (UTCTimestamp seconds, setMarkers migration, nonexistent timezone option, per-bar colors, SSR) with copy-pasteable v5-correct code — clear marginal value over manual doc lookup or stale snippets. Static cap of 7: examples unexecuted, direct-usability of outputs unverified against representative results.
The skill itself mandates local-typings-first verification with copyable rg/node commands — auditable primary material; repo CI and test suites corroborate maintenance quality. However, no independent third-party reproduction of the skill itself, and many routing-table doc paths are unconfirmed in this evidence, so not above 5.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →