SuperAPI AI Directory

docs-and-website-sync

Writing & research

View repository

Updated 10-15 17:21

About this skill

docs-and-website-sync

Purpose

Based on the repository's actual capabilities, update website、docs、README、skills, scaffolding AI guidance, and packaged documentation to prevent public entry points from lagging behind the implementation.

When to use

  • The user requests to "update website/docs/skills based on the existing code".
  • weapp-vite、wevu、weapp-ide-cli or the scaffolding has added configuration, CLI commands, or AI workflows.
  • Drift occurs among website configuration pages, AI guides, package descriptions, and skills.
  • Add or adjust these public capabilities:
    • weapp.autoRoutes / routeRules / layout
    • .weapp-vite supported files and wv prepare
    • forwardConsole / mcp init|print|doctor
    • wv screenshot / wv compare / wv ide logs
    • wv analyze --markdown|--report pr|--budget-check|--hmr-profile and weapp.analyze
    • web runtime / lib mode / multi-platform / npm / subpackages / worker
    • React 19 / JSX/TSX / mpcore headless parity
    • create-weapp-vite AI skills installation and AGENTS.md

Not applicable scenarios

This skill focuses on "synchronizing external documentation and skill entry points".

  • Project configuration, build strategy, CLI distribution: use weapp-vite-best-practices。
  • .vue Macros, templates, and SFC compatibility: use weapp-vite-vue-sfc-best-practices。
  • wevu Runtime semantics: use wevu-best-practices。
  • DevTools runtime e2e: use weapp-devtools-e2e-best-practices。
  • React runtime, render mode, and component bridge: use weapp-vite-react-best-practices。

Core workflow

  1. Check the source of truth first; do not copy from one another based on outdated wording:
    • packages/weapp-vite/src/types/config/**
    • packages/weapp-vite/src/cli/**
    • packages/weapp-ide-cli/src/cli/**
    • packages-runtime/react/src/**
    • packages-runtime/wevu/src/**
    • mpcore/packages/**
    • packages/weapp-vite/docs/README.md
    • packages/weapp-vite/docs/mcp.md
    • packages/weapp-vite/docs/volar.md
    • packages/weapp-vite/docs/define-config-overloads.md
    • packages/weapp-vite/docs/packaged/**
    • packages/weapp-vite/scripts/sync-package-docs.mjs
    • packages/create-weapp-vite/src/agents.ts
    • packages/create-weapp-vite/src/skills.ts
    • skills/*/SKILL.md
  2. Establish a "capability change -> entry page" mapping, then modify uniformly:
    • Configuration changes:website/config/**
    • CLI / AI / MCP:website/guide/**、website/packages/**、website/.vitepress/components/AiLearningPage.vue、packages/weapp-vite/README.md
    • Scaffolding and AI constraints:website/packages/create-weapp-vite.md、packages/create-weapp-vite/src/agents.ts
    • Runtime and framework:website/config/vue.md、website/config/wevu.md、website/packages/web.md
    • React:website/config/react.md、docs/integration/react.md, React templates and @weapp-vite/react README
    • Multi-platform and headless:website/guide/multi-platform.md、mpcore package docs、provider-compatible e2e
  3. Synchronize the AI contract wording:
    • First read the root AGENTS.md
    • then read node_modules/weapp-vite/dist/docs/*.md
    • MCP 接入 -> wv mcp init|print|doctor <codex|claude-code|cursor>
    • 运行时检查 -> weapp_devtools_* / weapp_runtime_* tools
    • 截图 -> wv screenshot / take_weapp_screenshot
    • 截图对比 -> wv compare / compare_weapp_screenshot
    • 日志 -> wv ide logs --open
    • 分析报告 -> wv analyze --markdown|--report pr|--budget-check|--hmr-profile
    • .weapp-vite supported files -> wv prepare
  4. Generated assets are refreshed only through scripts or builds; do not edit them manually:
    • website/public/llms-index.json
    • website/public/seo-quality-report.json
  5. If a skill is added, deleted, or merged, also update:
    • website/guide/skills.md
    • website/guide/ai-workflows.md
    • website/guide/ai.md
    • website/.vitepress/components/AiLearningPage.vue
    • skills/skill-trigger-regression-checklist.md
    • skills/scripts/score-skill-trigger-regression.mjs
    • CLAUDE.md and scaffolding AGENTS.md templates
  6. Verify with the minimal scope:
    • pnpm exec eslint --no-warn-ignored skills/*/SKILL.md
    • When involving website/** , run pnpm --filter website-weapp-vite build
    • When involving skill triggering/scoring, also run pnpm skills:check:yaml、pnpm skills:score:json

Capability topic mapping

  • HMR: use packages/weapp-vite/docs/packaged/weapp-config.md、website/config/hmr.md and real e2e as the source of truth.
  • Plugins: use pluginRoot types,website/guide/plugin.md, template AGENTS, and dist-plugin acceptance as the source of truth.
  • router/Web runtime: verify packages-runtime/wevu/docs/router-*.md and website/guide/web-runtime-*.mdseparately; do not merge them into browser semantics.
  • native AST/HMR performance: synchronize profile, fallback, and correctness conclusions, not just multiplier marketing claims.
  • React/Wevu JSX: clarify project-level ownership, WeChat platform limitations, and bridge boundaries; do not conflate the two JSX compilation chains.
  • mpcore parity: synchronize provider selection, the three-layer override, and the incremental rule that "actual DevTools behavior is authoritative".
  • skill metadata: every public skill should have agents/openai.yaml, and the trigger regression checklist should include primary scenarios, boundary scenarios, and conflict scenarios.

Constraints

  • Do not skip source verification and directly modify the wording.
  • Do not modify only a single entry point and miss README, website, packaged docs, or skills.
  • Do not manually maintain generated assets.
  • Do not describe the Mini Program runtime screenshot capability as a generalized browser screenshot.
  • Do not ignore the scaffold-generated AGENTS.md and the local dist/docs; they are part of the current product contract.

Output

When applying this skill, the output must include:

  • Capability change points.
  • List of affected entry points.
  • Specific modified files.
  • Verification commands.
  • If there are generated assets, explain how to refresh them.

Completion Markers

  • The configuration page, CLI page, AI guide, package description, skills, and the actual implementation are consistent.
  • wv / weapp-vite dual commands,prepare、mcp init|print|doctor、forwardConsole、analyze、screenshot/compare/ide logs has no documentation drift.
  • create-weapp-vite 's AI skills installation is consistent with the AGENTS.md template description.
  • The priority of the local bundled documentation dist/docs has been clearly stated.

Reference Materials

  • references/docs-sync-checklist.md