Container Queries — fixed Baseline status: corrected from "Widely available since February 2023" to "Widely available since August 2025" (February 2023 was Newly available)
HTML Sanitizer API — updated to reflect redesigned API: corrected Sanitizer constructor option names and browser support versions per current MDN documentation
Performance & Core Web Vitals — fixed INP metric description: clarified it reports near-worst-case (98th percentile for 50+ interactions), not absolute worst-case
Visual Regression Testing with Vitest — new standalone recipe (split out of Testing Strategy) covering Vitest 4 provider packages, locator-based toMatchScreenshot, baseline workflow, and CI/OS consistency caveats; Testing Strategy now carries a short summary linking the recipe
Firebase — new Backend Services tech stack category with Firebase as the recommended BaaS: rationale, alternatives, pricing tiers, and when-not-to-use guidance
Token Consumption Optimization — new guide covering strategies to reduce AI tool token consumption, ordered by impact: model selection, thinking effort, multi-agent model routing, context management, conversation compaction, and usage analysis; expanded subagent section with guidance on delegating file search and reading to small utility models
npm Workspaces — new learning page covering workspace setup, command targeting, cross-workspace linking, the semver-match registry fallback, and limitations vs pnpm/Yarn
Publishing from a Monorepo — corrected workspace protocol support: npm does not support the workspace: protocol (pnpm and Yarn Berry do)
AI Assistant Chat cookbook — added note that TanStack Virtual is under evaluation for replacing scroll anchor flags in Zustand, with link to the chat primitives blog post
AI-Assisted Development — added LLM Wiki blog series links to Further Reading (LLM Wiki pattern, Aliz Web Hub implementation, local models, and OKF v0.1 formal spec)
Triage — restructured into folder with dedicated AI Tooling sub-section; added Vercel AI SDK and MCP Apps evaluation pages
MCP Servers — added section on the 2026-07-28 specification release candidate: stateless core redesign, MCP Apps, Tasks extension, observability headers, and deprecated features (Roots, Sampling, Logging) with July 2027 sunset
MCP Servers — added MDN Web Docs MCP Server section: remote HTTP server (https://mcp.mdn.mozilla.net/) providing mcp_mdn_search, mcp_mdn_get-doc, and mcp_mdn_get-compat tools; already configured in the project's .vscode/mcp.json
Security Overview — corrected SameSite cookie default: SameSite=Lax is the default only in Chromium-based browsers (Chrome 80+, Edge 86); Firefox and Safari do not yet default to Lax
HTML Sanitizer API — corrected experimental browser support to Chrome 146+ (was Chrome 124+)
URLPattern — corrected Safari support to Safari 26+ (was Safari 16.4+)
Container Queries — updated CSS container style queries browser support to include Safari 18+ and Firefox 151+ alongside Chrome/Edge 111+
Zod — added LLM structured output as a When to use case: constraining model responses with Zod 4's z.toJSONSchema() across OpenAI, Anthropic, and Gemini
Triage — added Pulumi (TypeScript SDK) for evaluation as an IaC solution for Node.js and frontend projects
Copilot SDK: What You Can Build With It — practical overview of the Copilot SDK as a programmable agent runtime: usage ideas, comparison with LangGraph/CrewAI, cookbook patterns, and decision guidance
Copilot Automations: Your AI Team Member That Never Sleeps 🤖⏰ — blog post covering scheduled and event-driven Copilot Automations: use cases (daily docs updates, micro-refactors, issue triage, PR reports), comparison with GitHub Actions, and configuration examples. Updated to match official GitHub docs: corrected plan availability, trigger types, setup steps, and added secrets/privacy guidance
The LLM Wiki Now Has a Formal Spec — OKF v0.1 📐 — analysis of the Open Knowledge Format v0.1, a vendor-neutral spec that formalizes the LLM Wiki pattern into an interoperable standard for AI-consumable knowledge
Added web-api-fact-check.instructions.md — Copilot instruction that enforces fact-checking Web API, CSS, and HTTP references against MDN Web Docs before publishing
Pinned GitHub Actions to full-length commit SHAs for supply chain security (docs-build workflow)
Removed the leftover Docusaurus tutorial boilerplate (docs-tutorial/) from the docs sidebar
Removed Docusaurus scaffold blog posts and sample author entries from blog/authors.yml
Added .vscode/mcp.json — MDN MCP server configured for the repository; connects VS Code Copilot to https://mcp.mdn.mozilla.net/ for web platform documentation and compatibility lookups
Updated web-api-fact-check.instructions.md — integrated MDN MCP server tools (mcp_mdn_search, mcp_mdn_get-doc, mcp_mdn_get-compat) as the primary verification method; manual MDN URL lookups become the fallback when MCP is unavailable
AI-Assisted Development — complete new section covering the full spectrum from autocomplete to multi-agent pipelines, including agent skills, prompt engineering, MCP servers, and vibe coding
UI/UX Design — new section with laws of UX, visual design, accessibility, UI patterns, and design tools
Security — added section with prototype pollution docs and OWASP-based npm security checklist
Tech Stack — the curated set of recommended tools for Aliz frontend projects, covering Build & Dev (Vite, Vitest, Playwright), Core (React, TypeScript), Data Visualization (Recharts, ECharts), State & Data (TanStack Query), and Styling & UI (Tailwind CSS, shadcn/ui, MUI)
Rendering — architecture strategies and trade-offs (SSR, SSG, CSR, ISR)
Testing — three-layer testing strategy (unit, visual regression, E2E) with Vitest and Playwright coverage
Requirement Engineering — expanded with B2B/B2C considerations, browser support, i18n, accessibility, and performance metrics
Estimation — new Cookbooks section with React SPA recipe and complexity factors reference