AI Decision System
Replayable, offline-verifiable records for AI-assisted go/no-go decisions: experiment promotions, typed model answers and human-approved actions.
SvelteKit 앱 코드를 배포 전 검사해 SEO·성능·정확성·보안·구조·접근성을 확인
Your SvelteKit app's code health, checked before it ships. Statically, no browser. Covers SEO, Performance, Correctness, Security, Architecture, and Accessibility.
개발 도구TypeScript · JavaScriptGitHub @oekazuma온라인운영자 미확인
써 보셨어요? 한 번만 눌러 주세요. 로그인은 필요 없어요.
Your SvelteKit app's code health, checked before it ships. Statically, no browser — SEO, Performance, Correctness, Security, Architecture, Accessibility.
저장소 README에서 가져옴
README 전체를 불러오는 중입니다. 아래는 저장된 발췌입니다.
Your SvelteKit app's code health, checked before it ships. Statically, no browser.
SEO · Performance · Correctness · Security · Architecture · Accessibility
Pre-1.0. Not recommended for production use yet. Development is moving fast and aggressively, driven at the maintainer's discretion until 1.0: APIs, rule IDs, scoring, and output formats can change at any time, including breaking changes between minor releases. Relying on it in critical pipelines is discouraged until 1.0.README 전문 보기 ↗npx svelte-vitals@latest
https://github.com/user-attachments/assets/0f2a6f24-eed1-4380-be9a-b09e728c4208
[!WARNING] Pre-1.0. Not recommended for production use yet. Development is moving fast and aggressively, driven at the maintainer's discretion until
1.0: APIs, rule IDs, scoring, and output formats can change at any time, including breaking changes between minor releases. Relying on it in critical pipelines is discouraged until1.0.
Six categories, each scored independently and rolled into a single weighted Health score: SEO, Performance, Correctness, Security, Architecture, Accessibility. → Health score
console, json, agent (a Markdown remediation document an AI agent can act on directly), sarif, github, md (a plain Markdown report), and html (a self-contained, shareable report). The agent reporter auto-selects inside AI-agent harnesses (e.g. Claude Code); github auto-selects under GitHub Actions. → Reporterssvelte-vitals ci install scaffolds a full GitHub Actions workflow around @svelte-vitals/action, with a job summary and a sticky PR comment. → Reporters, CI integration/__svelte-vitals/ during vite dev, on by default: whole-project coverage from startup, refined to real rendered values as you browse, with a copy-to-clipboard AI-agent prompt on every finding. → Live dashboard@svelte-vitals/vite): build-time analysis of the prerendered <head>; library-agnostic and exact. → Plugin modesvelte-vitals docs list / docs show <name> print the guides from inside the CLI, so they match the installed version and need no network; svelte-vitals explain --list / explain <rule-id> do the same for the rule set. docs list and both forms of explain take --json; docs show prints the topic as Markdown. → CLI reference/setup-svelte-vitals, /improve-svelte, and /svelte-vitals slash-command skills for Claude Code, Cursor, and Codex: a first-run pass that derives the config file from what the project already declares, a project-wide, evidence-ranked improvement roadmap, and rule knowledge up front. Install with npx skills add oekazuma/svelte-vitals (skills.sh). → Agent SkillsTwo packages you'll use directly, the CLI and the Vite plugin, both built on the shared @svelte-vitals/core rule engine. → Choosing a package
→ Getting started for installation, your first run, and exit codes for CI gating.
See CONTRIBUTING.md for development setup, commands, and the release process.