[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7ASTH1B7G6H9ETZPV1PVDME":3,"skill-same:01M7ASTH1B7G6H9ETZPV1PVDME":48},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":31,"ownerDescription":32,"updatedAt":33,"install":34,"resources":47},"01M7ASTH1B7G6H9ETZPV1PVDME","browser-engine-architecture-and-networking",null,"Use when working on or reasoning about browser architecture: the anatomy of a browser (the router for the whole web-browser-development reference), who actually renders the web (Blink, WebKit, Gecko, Servo, Ladybird) and what it costs to build one, what a browser is besides an engine, why multiple processes, Site Isolation, sandboxing per platform, IPC design, and the networking stack — HTTP\u002F1.1, HTTP\u002F2, HTTP\u002F3 and QUIC, TLS, DNS, and caching.","the-vibey-project\u002Fvibey","the-vibey-project","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F329298758?v=4","01M7AJHGZ9722CTJ4JNQZVRH0Q","the vibey project","01M7AJHGZVMG2K7TGXQ6XFHVG3",5,"MIT","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fweb-browser-development\u002Fskills\u002Fbrowser-engine-architecture-and-networking\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: browser-engine-architecture-and-networking\ndescription: \"Use when working on or reasoning about browser architecture: the anatomy of a browser (the router for the whole web-browser-development reference), who actually renders the web (Blink, WebKit, Gecko, Servo, Ladybird) and what it costs to build one, what a browser is besides an engine, why multiple processes, Site Isolation, sandboxing per platform, IPC design, and the networking stack — HTTP\u002F1.1, HTTP\u002F2, HTTP\u002F3 and QUIC, TLS, DNS, and caching.\"\n---\n\n# Browser Development: The Engine Landscape, Process Model and Sandboxing, and Networking\n\n> **Part 1 of 5** of the *Web Browser Development* reference (plugin `web-browser-development`), covering §0–§3. Sibling skills: `browser-rendering-pipeline` (§4–§7), `browser-security-and-privacy` (§8–§9), `browser-extensions-platform-and-standards` (§10–§14), `browser-development-reference` (§15–§20). Section numbers are shared across the set; a reference written as §N → `skill` points into that sibling skill.\n>\n> **Currency:** Verified August 2026. See §17 → `browser-development-reference` for the currency snapshot and what goes stale first.\n\n> **How to read this.** Reference, not tutorial. Sections are independent. Three markers:\n> - **[DURABLE]** — architecture, algorithms, or a constraint every engine has\n>   independently arrived at. Does not expire.\n> - **[ENGINE]** — specific to Blink\u002FChromium, Gecko, WebKit, or Servo\u002FLadybird. Verify\n>   against that project's source or docs.\n> - **[CONTESTED]** — the vendors genuinely disagree, often because their business models\n>   differ. Both cases given, including the commercial motive where it's load-bearing.\n>\n> **⚠️ GOTCHA** boxes mark the mistakes that produce security holes, jank, or\n> compatibility breakage that can never be undone.\n>\n> **The two framings that organize everything below:**\n> 1. **A browser is a hostile-input execution environment that must never say no.** It\n>    runs untrusted code from anyone, on documents that are frequently malformed, and it\n>    is not permitted to refuse or crash. Every architectural decision — the multi-process\n>    model, the error-tolerant parser, the sandbox — descends from that.\n> 2. **The web's compatibility constraint is stronger than any other platform's.** You\n>    cannot break existing sites. Not the well-written ones, not the abandoned ones. This\n>    is why the platform accretes and rarely subtracts, and why \"just fix it properly\"\n>    is almost never available.\n\n---\n\n## §0. Routing\n\n### 0.1 The anatomy\n\n```\n┌────────────────────────────────────────────────────────────────────────┐\n│ BROWSER PROCESS (privileged)                                           │\n│   UI · tab\u002Fwindow management · profile & storage · permissions         │\n│   navigation & session history · extension host · IPC broker           │\n└───────┬─────────────────┬──────────────────┬───────────────────────────┘\n        │ IPC             │ IPC              │ IPC\n┌───────▼──────┐  ┌───────▼────────┐  ┌──────▼───────┐  ┌───────────────┐\n│ RENDERER     │  │ NETWORK        │  │ GPU          │  │ UTILITY       │\n│ (sandboxed,  │  │ SERVICE        │  │ PROCESS      │  │ (audio, media │\n│  one per     │  │ HTTP\u002FTLS\u002Fcache │  │ raster, draw │  │  codecs, data │\n│  site)       │  │ cookies, DNS   │  │ compositing  │  │  decoders)    │\n│              │  └────────────────┘  └──────────────┘  └───────────────┘\n│  ┌─────────────────────────────────────────────┐\n│  │ HTML parser → DOM                            │\n│  │ CSS parser → CSSOM → style resolution        │\n│  │ LAYOUT → PAINT (display lists) → COMMIT      │\n│  │ JS engine (heap, JIT) · event loop            │\n│  │ compositor thread (scroll, transform anims)   │\n│  └─────────────────────────────────────────────┘\n└──────────────┘\n```\n\n**[DURABLE] The single most important architectural fact: the renderer is assumed to be\ncompromised.** It is sandboxed, it has no filesystem or network access of its own, and\nevery privileged operation goes through an IPC message that the browser process must\nindependently validate. **Never trust a renderer's claim about its own origin** — that is\nthe whole ballgame (§8.4 → `browser-security-and-privacy`).\n\n### 0.2 The question router\n\n| Asked about... | Go to |\n|---|---|\n| Engine landscape, who builds what, why it matters | §1 |\n| Process model, IPC, sandboxing, site isolation | §2 |\n| Networking, HTTP, TLS, caching, DNS | §3 |\n| HTML parsing, DOM, error recovery | §4 → `browser-rendering-pipeline` |\n| CSS: cascade, selectors, style resolution, invalidation | §5 → `browser-rendering-pipeline` |\n| Layout algorithms and fragmentation | §6.1 → `browser-rendering-pipeline` |\n| Paint, raster, compositing, GPU | §6.2 → `browser-rendering-pipeline`–6.4 |\n| JS engine integration, bindings, GC interaction | §7 → `browser-rendering-pipeline` |\n| Event loop, scheduling, rendering lifecycle | §7.3 → `browser-rendering-pipeline` |\n| Security: SOP, CORS, CSP, XSS, Spectre, memory safety | §8 → `browser-security-and-privacy` |\n| Privacy, tracking, cookies, fingerprinting | §9 → `browser-security-and-privacy` |\n| Extensions | §10 → `browser-extensions-platform-and-standards` |\n| Storage, media, devices, capability APIs | §11 → `browser-extensions-platform-and-standards` |\n| Accessibility | §12 → `browser-extensions-platform-and-standards` |\n| Standards, interop, compatibility | §13 → `browser-extensions-platform-and-standards` |\n| Testing, telemetry, shipping | §14 → `browser-extensions-platform-and-standards` |\n| \"Don't do this\" | §15 → `browser-development-reference` |\n| \"Which approach is better?\" | §16 → `browser-development-reference` (contested) |\n| \"Is this still current?\" | §17 → `browser-development-reference` |\n| Books, docs, people | §18 → `browser-development-reference` |\n\n---\n\n## §1. The Engine Landscape\n\n### 1.1 Who actually renders the web\n\n**[VERSIONED — as of 2026, three engines carry essentially all of it:]**\n\n| Engine | Share (desktop+mobile+tablet, ~May 2026) | Ships in |\n|---|---|---|\n| **Blink** (Chromium) | **~81%** | Chrome, Edge, Brave, Arc, Opera, Vivaldi, Samsung Internet, and nearly every Electron app |\n| **WebKit** (Apple) | **~14%** | Safari — and, **outside the EU**, *every* iOS browser, because the App Store mandated WebKit (Chrome\u002FFirefox\u002FEdge on iOS are WebKit wrappers) |\n| **Gecko** (Mozilla) | **~3%** | Firefox and derivatives. The last major engine that is neither Chromium nor Apple |\n\nThat's ~99%. The remaining ~1% is Pale Moon's Goanna, Ekioh's commercial GPU-accelerated\nFlow, and the two ground-up rewrites below.\n\n**The two new engines, and their honest status [VERSIONED]:**\n- **Servo** — originated at Mozilla as a Rust, parallel, embeddable engine co-developed\n  with Rust itself (2012–2020); donated to the Linux Foundation, now under **Linux\n  Foundation Europe**. Modular; several components (Stylo, WebRender) were upstreamed into\n  Firefox and are in production there. Positioned as an **embeddable webview**, not a\n  browser.\n- **Ladybird** — Andreas Kling's independent C++ engine (**LibWeb** + **LibJS**), spun out\n  of SerenityOS, backed by the non-profit Ladybird Browser Initiative (incorporated 2024),\n  funded by private sponsorship including GitHub and Shopify founders and Cloudflare, with\n  roughly 7–10 full-time engineers. **No code from Blink, WebKit, or Gecko.** Explicit\n  policy of never taking search-default money. **First Alpha targeted 2026 for Linux and\n  macOS**, aimed at developers and early adopters; **beta expected 2027, general stable\n  2028.** Roadmap items include moving style and layout to Rust, sandboxing, GPU isolation,\n  and WebAssembly GC — note that those are *roadmap*, not shipped.\n\n**[DURABLE] Why the count matters more than the market share.** With one engine at 81%,\n\"works in Chrome\" becomes the de facto standard regardless of what the spec says, and a\nsingle vendor's product decisions become web-wide policy. Every argument in §9 → `browser-security-and-privacy`, §10 → `browser-extensions-platform-and-standards`, and\n§13 → `browser-extensions-platform-and-standards` is downstream of this.\n\n### 1.2 What it costs to build one\n\n**[DURABLE] A browser engine is one of the largest artifacts in commercial software** —\ntens of millions of lines, decades of accumulated compatibility behaviour, and a security\nsurface that attracts state-level attackers. The realistic assessment:\n- **Nobody has built a competitive from-scratch engine in twenty years** and shipped it to\n  general users. Ladybird's 2026-alpha\u002F2028-stable timeline with ~10 full-time engineers is\n  the current experiment in whether that's still possible.\n- The hard part is **not** the parts that are fun (parser, layout algorithms). It's the\n  long tail: compatibility with two decades of quirks, media codecs and DRM, accessibility,\n  the security architecture, and the sheer surface area of the platform.\n- **Forking is the rational choice and is why the landscape looks like it does.** Every\n  \"new browser\" of the last decade is a Chromium shell with different UI and policy.\n\n### 1.3 What a browser is besides an engine\n\nProduct surface that is genuinely half the work: profiles and sync, password and payment\nmanagement, downloads, bookmarks and history, autofill, the update mechanism (a\n**security-critical** always-on channel), enterprise policy, crash reporting and telemetry,\nDevTools, and the extension platform.\n\n---\n\n## §2. Process Model, IPC, and Sandboxing\n\n### 2.1 Why multiple processes\n\n**[DURABLE]** Chrome's 2008 multi-process design solved three problems at once and every\nengine has since converged on it: **stability** (a renderer crash kills a tab, not the\nbrowser), **performance** (parallelism across cores; one page's main thread can't block\nanother's), and — the one that turned out to matter most — **security** (renderers run in\na restricted sandbox with no direct filesystem, network, or device access).\n\n**The Chromium terminology, precisely, because people conflate these:**\n- **Multi-process architecture** — the broad design choice of separate OS processes.\n- **Site Isolation** — the stricter *policy* that a renderer process is locked to a single\n  site or origin, including for iframes.\n- **Sandboxing** — the OS-level restriction layer limiting what a compromised process can\n  do *after* code execution.\n\n### 2.2 Site Isolation\n\n**[ENGINE, and the reference design]** Chrome enabled Site Isolation by default in Chrome\n67, motivated directly by **Spectre** (§8.6 → `browser-security-and-privacy`). Its two halves:\n1. **Locked renderer processes** — a renderer may contain documents and workers from only\n   one site or origin, *even in iframes*. Getting this right required solving genuinely\n   hard cases: `srcdoc` URLs, `data:` URLs, base URLs, and sandboxed frames (Chrome enabled\n   **isolated sandboxed frames** by default in 2024, adding a process boundary between an\n   origin and the untrustworthy content it hosts).\n2. **Browser-enforced restrictions** — the privileged browser process validates every IPC\n   message and refuses cross-site data requests (`ChildProcessSecurityPolicy::\n   CanAccessDataForOrigin` is the canonical check). This is what stops a *fully compromised*\n   renderer from simply asking for another site's cookies.\n\n**Firefox's equivalent is Project Fission**, which isolates at the **origin** boundary\nrather than the site boundary — stricter than Chrome's default, because Chrome's site\nboundary does not separate `mail.example.com` from `pay.example.com`.\n\n**[DURABLE] The security payoff, stated precisely:** under Site Isolation, an attacker with\nfull remote code execution inside a renderer still cannot read another site's DOM, cookies,\nor JS heap, because **those objects are simply not in that process's address space**. To\nreach them the attacker must chain a second exploit — a browser-process privilege\nescalation or a kernel sandbox escape. That's the whole design.\n\n> **⚠️ GOTCHA — the cost is memory and process count, and it binds hardest on mobile.**\n> Chromium's own security documentation is unusually candid: *\"we are reaching the limits\n> of sandboxing and site isolation. A key limitation is that the process is the smallest\n> unit of isolation, but processes are not cheap. Especially on Android, using more\n> processes impacts device health overall: background activities get killed with far\n> greater frequency.\"* It also notes that processes still share information about multiple\n> sites — the network service is one large C++ component parsing complex input from\n> anyone on the network. **Process isolation has a ceiling, and the industry has hit it.**\n\n### 2.3 Sandboxing, per platform\n\n| Platform | Mechanism |\n|---|---|\n| **Windows** | Restricted token + job object + alternate desktop + **AppContainer** + **Win32k lockdown** (blocking direct access to the win32k syscall surface, historically a huge kernel attack surface) |\n| **macOS** | Seatbelt (`sandbox_init`) profiles |\n| **Linux** | **seccomp-bpf** syscall filter + user namespaces + `setuid` sandbox (legacy) |\n| **Android** | Process sandbox reinforced by **SELinux** policy; isolated processes |\n| **iOS** | System sandbox; JIT requires special entitlement — hence the WebKit mandate historically |\n\n**[DURABLE] Renderers must be denied filesystem, network, and device access.** All of it\nroutes through brokered IPC to the browser process, which validates. Any capability you\nhand directly to the renderer is a capability the attacker gets.\n\n### 2.4 IPC design\n\n**[DURABLE] The IPC layer is a security boundary, and it must be treated like a network\nprotocol from a hostile peer.** The rules:\n- **Validate everything in the privileged process.** Never trust a size, an index, a\n  handle, or — especially — an *origin* claimed by a renderer.\n- **Capability-style interfaces** (Chromium's **Mojo**) beat giant switch statements on\n  message IDs: a renderer holds a pipe to a specific service with a specific interface,\n  rather than the ability to send any message.\n- **Fuzz the IPC surface.** It is the highest-value fuzzing target in the browser.\n- Mind the **serialization** cost: IPC on the rendering hot path shows up directly in\n  frame time.\n\n---\n\n## §3. Networking\n\n### 3.1 The stack\n\n```\nURL parse (WHATWG URL — the spec exists because everyone did this differently)\n  → HSTS check → scheme handling\n  → DNS (system, DoH, DoT) → Happy Eyeballs (race IPv4\u002FIPv6, prefer fast path)\n  → TCP + TLS 1.3, or QUIC (UDP) for HTTP\u002F3\n  → connection pool \u002F coalescing\n  → HTTP\u002F1.1 · HTTP\u002F2 (multiplex) · HTTP\u002F3 (QUIC, no head-of-line blocking)\n  → response: caching, decompression (gzip\u002Fbr\u002Fzstd), MIME sniffing\n  → hand to renderer as a stream\n```\n\n**Points that matter for implementers:**\n- **HTTP\u002F2** multiplexes over one TCP connection, eliminating per-request connections —\n  but a lost packet stalls *all* streams (TCP head-of-line blocking).\n- **HTTP\u002F3 \u002F QUIC** moves to UDP with per-stream loss recovery, fixing that, plus\n  0-RTT resumption and connection migration across network changes. QUIC is implemented in\n  userspace, which is why browsers ship their own.\n- **Connection coalescing**: reusing one connection for multiple hosts that resolve to the\n  same IP with a covering certificate. A real performance win and a subtle correctness trap.\n- **HTTP caching is a specification, not a heuristic** — `Cache-Control`, `ETag`,\n  `Last-Modified`, `Vary`, revalidation, and the freshness lifetime rules. ⚠️ **Cache\n  partitioning by top-level site is now standard** (§9.2 → `browser-security-and-privacy`) and changed the performance\n  calculus of shared CDN resources permanently.\n- **MIME sniffing** exists because servers lie. It is also a security hazard —\n  `X-Content-Type-Options: nosniff` exists to turn it off, and sniffing a response into an\n  executable type is a classic XSS vector.\n- **Preload\u002Fpreconnect\u002Fprefetch\u002Fpriority hints** — the browser's scheduling of what to\n  fetch when is a large part of real-world page speed.\n\n> **⚠️ GOTCHA — the network service parses hostile input from everyone, in C++, for all\n> sites at once.** Chromium names this explicitly as a residual risk (§2.2). Whatever\n> your architecture, that component deserves memory-safe implementation, aggressive\n> fuzzing, and its own sandbox.\n","# Browser Development: The Engine Landscape, Process Model and Sandboxing, and Networking\n\n> **Part 1 of 5** of the *Web Browser Development* reference (plugin `web-browser-development`), covering §0–§3. Sibling skills: `browser-rendering-pipeline` (§4–§7), `browser-security-and-privacy` (§8–§9), `browser-extensions-platform-and-standards` (§10–§14), `browser-development-reference` (§15–§20). Section numbers are shared across the set; a reference written as §N → `skill` points into that sibling skill.\n>\n> **Currency:** Verified August 2026. See §17 → `browser-development-reference` for the currency snapshot and what goes stale first.\n\n> **How to read this.** Reference, not tutorial. Sections are independent. Three markers:\n> - **[DURABLE]** — architecture, algorithms, or a constraint every engine has\n>   independently arrived at. Does not expire.\n> - **[ENGINE]** — specific to Blink\u002FChromium, Gecko, WebKit, or Servo\u002FLadybird. Verify\n>   against that project's source or docs.\n> - **[CONTESTED]** — the vendors genuinely disagree, often because their business models\n>   differ. Both cases given, including the commercial motive where it's load-bearing.\n>\n> **⚠️ GOTCHA** boxes mark the mistakes that produce security holes, jank, or\n> compatibility breakage that can never be undone.\n>\n> **The two framings that organize everything below:**\n> 1. **A browser is a hostile-input execution environment that must never say no.** It\n>    runs untrusted code from anyone, on documents that are frequently malformed, and it\n>    is not permitted to refuse or crash. Every architectural decision — the multi-process\n>    model, the error-tolerant parser, the sandbox — descends from that.\n> 2. **The web's compatibility constraint is stronger than any other platform's.** You\n>    cannot break existing sites. Not the well-written ones, not the abandoned ones. This\n>    is why the platform accretes and rarely subtracts, and why \"just fix it properly\"\n>    is almost never available.\n\n---\n\n## §0. Routing\n\n### 0.1 The anatomy\n\n```\n┌────────────────────────────────────────────────────────────────────────┐\n│ BROWSER PROCESS (privileged)                                           │\n│   UI · tab\u002Fwindow management · profile & storage · permissions         │\n│   navigation & session history · extension host · IPC broker           │\n└───────┬─────────────────┬──────────────────┬───────────────────────────┘\n        │ IPC             │ IPC              │ IPC\n┌───────▼──────┐  ┌───────▼────────┐  ┌──────▼───────┐  ┌───────────────┐\n│ RENDERER     │  │ NETWORK        │  │ GPU          │  │ UTILITY       │\n│ (sandboxed,  │  │ SERVICE        │  │ PROCESS      │  │ (audio, media │\n│  one per     │  │ HTTP\u002FTLS\u002Fcache │  │ raster, draw │  │  codecs, data │\n│  site)       │  │ cookies, DNS   │  │ compositing  │  │  decoders)    │\n│              │  └────────────────┘  └──────────────┘  └───────────────┘\n│  ┌─────────────────────────────────────────────┐\n│  │ HTML parser → DOM                            │\n│  │ CSS parser → CSSOM → style resolution        │\n│  │ LAYOUT → PAINT (display lists) → COMMIT      │\n│  │ JS engine (heap, JIT) · event loop            │\n│  │ compositor thread (scroll, transform anims)   │\n│  └─────────────────────────────────────────────┘\n└──────────────┘\n```\n\n**[DURABLE] The single most important architectural fact: the renderer is assumed to be\ncompromised.** It is sandboxed, it has no filesystem or network access of its own, and\nevery privileged operation goes through an IPC message that the browser process must\nindependently validate. **Never trust a renderer's claim about its own origin** — that is\nthe whole ballgame (§8.4 → `browser-security-and-privacy`).\n\n### 0.2 The question router\n\n| Asked about... | Go to |\n|---|---|\n| Engine landscape, who builds what, why it matters | §1 |\n| Process model, IPC, sandboxing, site isolation | §2 |\n| Networking, HTTP, TLS, caching, DNS | §3 |\n| HTML parsing, DOM, error recovery | §4 → `browser-rendering-pipeline` |\n| CSS: cascade, selectors, style resolution, invalidation | §5 → `browser-rendering-pipeline` |\n| Layout algorithms and fragmentation | §6.1 → `browser-rendering-pipeline` |\n| Paint, raster, compositing, GPU | §6.2 → `browser-rendering-pipeline`–6.4 |\n| JS engine integration, bindings, GC interaction | §7 → `browser-rendering-pipeline` |\n| Event loop, scheduling, rendering lifecycle | §7.3 → `browser-rendering-pipeline` |\n| Security: SOP, CORS, CSP, XSS, Spectre, memory safety | §8 → `browser-security-and-privacy` |\n| Privacy, tracking, cookies, fingerprinting | §9 → `browser-security-and-privacy` |\n| Extensions | §10 → `browser-extensions-platform-and-standards` |\n| Storage, media, devices, capability APIs | §11 → `browser-extensions-platform-and-standards` |\n| Accessibility | §12 → `browser-extensions-platform-and-standards` |\n| Standards, interop, compatibility | §13 → `browser-extensions-platform-and-standards` |\n| Testing, telemetry, shipping | §14 → `browser-extensions-platform-and-standards` |\n| \"Don't do this\" | §15 → `browser-development-reference` |\n| \"Which approach is better?\" | §16 → `browser-development-reference` (contested) |\n| \"Is this still current?\" | §17 → `browser-development-reference` |\n| Books, docs, people | §18 → `browser-development-reference` |\n\n---\n\n## §1. The Engine Landscape\n\n### 1.1 Who actually renders the web\n\n**[VERSIONED — as of 2026, three engines carry essentially all of it:]**\n\n| Engine | Share (desktop+mobile+tablet, ~May 2026) | Ships in |\n|---|---|---|\n| **Blink** (Chromium) | **~81%** | Chrome, Edge, Brave, Arc, Opera, Vivaldi, Samsung Internet, and nearly every Electron app |\n| **WebKit** (Apple) | **~14%** | Safari — and, **outside the EU**, *every* iOS browser, because the App Store mandated WebKit (Chrome\u002FFirefox\u002FEdge on iOS are WebKit wrappers) |\n| **Gecko** (Mozilla) | **~3%** | Firefox and derivatives. The last major engine that is neither Chromium nor Apple |\n\nThat's ~99%. The remaining ~1% is Pale Moon's Goanna, Ekioh's commercial GPU-accelerated\nFlow, and the two ground-up rewrites below.\n\n**The two new engines, and their honest status [VERSIONED]:**\n- **Servo** — originated at Mozilla as a Rust, parallel, embeddable engine co-developed\n  with Rust itself (2012–2020); donated to the Linux Foundation, now under **Linux\n  Foundation Europe**. Modular; several components (Stylo, WebRender) were upstreamed into\n  Firefox and are in production there. Positioned as an **embeddable webview**, not a\n  browser.\n- **Ladybird** — Andreas Kling's independent C++ engine (**LibWeb** + **LibJS**), spun out\n  of SerenityOS, backed by the non-profit Ladybird Browser Initiative (incorporated 2024),\n  funded by private sponsorship including GitHub and Shopify founders and Cloudflare, with\n  roughly 7–10 full-time engineers. **No code from Blink, WebKit, or Gecko.** Explicit\n  policy of never taking search-default money. **First Alpha targeted 2026 for Linux and\n  macOS**, aimed at developers and early adopters; **beta expected 2027, general stable\n  2028.** Roadmap items include moving style and layout to Rust, sandboxing, GPU isolation,\n  and WebAssembly GC — note that those are *roadmap*, not shipped.\n\n**[DURABLE] Why the count matters more than the market share.** With one engine at 81%,\n\"works in Chrome\" becomes the de facto standard regardless of what the spec says, and a\nsingle vendor's product decisions become web-wide policy. Every argument in §9 → `browser-security-and-privacy`, §10 → `browser-extensions-platform-and-standards`, and\n§13 → `browser-extensions-platform-and-standards` is downstream of this.\n\n### 1.2 What it costs to build one\n\n**[DURABLE] A browser engine is one of the largest artifacts in commercial software** —\ntens of millions of lines, decades of accumulated compatibility behaviour, and a security\nsurface that attracts state-level attackers. The realistic assessment:\n- **Nobody has built a competitive from-scratch engine in twenty years** and shipped it to\n  general users. Ladybird's 2026-alpha\u002F2028-stable timeline with ~10 full-time engineers is\n  the current experiment in whether that's still possible.\n- The hard part is **not** the parts that are fun (parser, layout algorithms). It's the\n  long tail: compatibility with two decades of quirks, media codecs and DRM, accessibility,\n  the security architecture, and the sheer surface area of the platform.\n- **Forking is the rational choice and is why the landscape looks like it does.** Every\n  \"new browser\" of the last decade is a Chromium shell with different UI and policy.\n\n### 1.3 What a browser is besides an engine\n\nProduct surface that is genuinely half the work: profiles and sync, password and payment\nmanagement, downloads, bookmarks and history, autofill, the update mechanism (a\n**security-critical** always-on channel), enterprise policy, crash reporting and telemetry,\nDevTools, and the extension platform.\n\n---\n\n## §2. Process Model, IPC, and Sandboxing\n\n### 2.1 Why multiple processes\n\n**[DURABLE]** Chrome's 2008 multi-process design solved three problems at once and every\nengine has since converged on it: **stability** (a renderer crash kills a tab, not the\nbrowser), **performance** (parallelism across cores; one page's main thread can't block\nanother's), and — the one that turned out to matter most — **security** (renderers run in\na restricted sandbox with no direct filesystem, network, or device access).\n\n**The Chromium terminology, precisely, because people conflate these:**\n- **Multi-process architecture** — the broad design choice of separate OS processes.\n- **Site Isolation** — the stricter *policy* that a renderer process is locked to a single\n  site or origin, including for iframes.\n- **Sandboxing** — the OS-level restriction layer limiting what a compromised process can\n  do *after* code execution.\n\n### 2.2 Site Isolation\n\n**[ENGINE, and the reference design]** Chrome enabled Site Isolation by default in Chrome\n67, motivated directly by **Spectre** (§8.6 → `browser-security-and-privacy`). Its two halves:\n1. **Locked renderer processes** — a renderer may contain documents and workers from only\n   one site or origin, *even in iframes*. Getting this right required solving genuinely\n   hard cases: `srcdoc` URLs, `data:` URLs, base URLs, and sandboxed frames (Chrome enabled\n   **isolated sandboxed frames** by default in 2024, adding a process boundary between an\n   origin and the untrustworthy content it hosts).\n2. **Browser-enforced restrictions** — the privileged browser process validates every IPC\n   message and refuses cross-site data requests (`ChildProcessSecurityPolicy::\n   CanAccessDataForOrigin` is the canonical check). This is what stops a *fully compromised*\n   renderer from simply asking for another site's cookies.\n\n**Firefox's equivalent is Project Fission**, which isolates at the **origin** boundary\nrather than the site boundary — stricter than Chrome's default, because Chrome's site\nboundary does not separate `mail.example.com` from `pay.example.com`.\n\n**[DURABLE] The security payoff, stated precisely:** under Site Isolation, an attacker with\nfull remote code execution inside a renderer still cannot read another site's DOM, cookies,\nor JS heap, because **those objects are simply not in that process's address space**. To\nreach them the attacker must chain a second exploit — a browser-process privilege\nescalation or a kernel sandbox escape. That's the whole design.\n\n> **⚠️ GOTCHA — the cost is memory and process count, and it binds hardest on mobile.**\n> Chromium's own security documentation is unusually candid: *\"we are reaching the limits\n> of sandboxing and site isolation. A key limitation is that the process is the smallest\n> unit of isolation, but processes are not cheap. Especially on Android, using more\n> processes impacts device health overall: background activities get killed with far\n> greater frequency.\"* It also notes that processes still share information about multiple\n> sites — the network service is one large C++ component parsing complex input from\n> anyone on the network. **Process isolation has a ceiling, and the industry has hit it.**\n\n### 2.3 Sandboxing, per platform\n\n| Platform | Mechanism |\n|---|---|\n| **Windows** | Restricted token + job object + alternate desktop + **AppContainer** + **Win32k lockdown** (blocking direct access to the win32k syscall surface, historically a huge kernel attack surface) |\n| **macOS** | Seatbelt (`sandbox_init`) profiles |\n| **Linux** | **seccomp-bpf** syscall filter + user namespaces + `setuid` sandbox (legacy) |\n| **Android** | Process sandbox reinforced by **SELinux** policy; isolated processes |\n| **iOS** | System sandbox; JIT requires special entitlement — hence the WebKit mandate historically |\n\n**[DURABLE] Renderers must be denied filesystem, network, and device access.** All of it\nroutes through brokered IPC to the browser process, which validates. Any capability you\nhand directly to the renderer is a capability the attacker gets.\n\n### 2.4 IPC design\n\n**[DURABLE] The IPC layer is a security boundary, and it must be treated like a network\nprotocol from a hostile peer.** The rules:\n- **Validate everything in the privileged process.** Never trust a size, an index, a\n  handle, or — especially — an *origin* claimed by a renderer.\n- **Capability-style interfaces** (Chromium's **Mojo**) beat giant switch statements on\n  message IDs: a renderer holds a pipe to a specific service with a specific interface,\n  rather than the ability to send any message.\n- **Fuzz the IPC surface.** It is the highest-value fuzzing target in the browser.\n- Mind the **serialization** cost: IPC on the rendering hot path shows up directly in\n  frame time.\n\n---\n\n## §3. Networking\n\n### 3.1 The stack\n\n```\nURL parse (WHATWG URL — the spec exists because everyone did this differently)\n  → HSTS check → scheme handling\n  → DNS (system, DoH, DoT) → Happy Eyeballs (race IPv4\u002FIPv6, prefer fast path)\n  → TCP + TLS 1.3, or QUIC (UDP) for HTTP\u002F3\n  → connection pool \u002F coalescing\n  → HTTP\u002F1.1 · HTTP\u002F2 (multiplex) · HTTP\u002F3 (QUIC, no head-of-line blocking)\n  → response: caching, decompression (gzip\u002Fbr\u002Fzstd), MIME sniffing\n  → hand to renderer as a stream\n```\n\n**Points that matter for implementers:**\n- **HTTP\u002F2** multiplexes over one TCP connection, eliminating per-request connections —\n  but a lost packet stalls *all* streams (TCP head-of-line blocking).\n- **HTTP\u002F3 \u002F QUIC** moves to UDP with per-stream loss recovery, fixing that, plus\n  0-RTT resumption and connection migration across network changes. QUIC is implemented in\n  userspace, which is why browsers ship their own.\n- **Connection coalescing**: reusing one connection for multiple hosts that resolve to the\n  same IP with a covering certificate. A real performance win and a subtle correctness trap.\n- **HTTP caching is a specification, not a heuristic** — `Cache-Control`, `ETag`,\n  `Last-Modified`, `Vary`, revalidation, and the freshness lifetime rules. ⚠️ **Cache\n  partitioning by top-level site is now standard** (§9.2 → `browser-security-and-privacy`) and changed the performance\n  calculus of shared CDN resources permanently.\n- **MIME sniffing** exists because servers lie. It is also a security hazard —\n  `X-Content-Type-Options: nosniff` exists to turn it off, and sniffing a response into an\n  executable type is a classic XSS vector.\n- **Preload\u002Fpreconnect\u002Fprefetch\u002Fpriority hints** — the browser's scheduling of what to\n  fetch when is a large part of real-world page speed.\n\n> **⚠️ GOTCHA — the network service parses hostile input from everyone, in C++, for all\n> sites at once.** Chromium names this explicitly as a residual risk (§2.2). Whatever\n> your architecture, that component deserves memory-safe implementation, aggressive\n> fuzzing, and its own sandbox.\n",{"name":5,"description":7},"develop","ae32736a1aabddabca5d2c96d665aa0ebc7d8700","src\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fweb-browser-development\u002Fskills\u002Fbrowser-engine-architecture-and-networking",3,"https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey","https:\u002F\u002Fthe-vibey-project.github.io\u002Fvibey\u002F","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project","https:\u002F\u002Fvibewithadam.matthewsteinberger.com","freedom for software engineers","2026-11-15 05:47:42",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add the-vibey-project\u002Fvibey --skill browser-engine-architecture-and-networking",{"tool":40,"command":41},"bunx","bunx skills add the-vibey-project\u002Fvibey --skill browser-engine-architecture-and-networking",{"tool":43,"command":44},"pnpm","pnpm dlx skills add the-vibey-project\u002Fvibey --skill browser-engine-architecture-and-networking","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fweb-browser-development\u002Fskills\u002Fbrowser-engine-architecture-and-networking\u002FSKILL.md",true,[],[49,56,63,70,77,84],{"skillNo":50,"name":51,"nameZh":6,"description":52,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":53,"categories":54,"keywords":55,"publishedAt":19},"01M7AT1FQJAE1FQYAMDTF490Y4","white-paper-writing","Use when writing white papers, technical reports, policy documents, thought leadership pieces, or any long-form authoritative document. Triggers on requests to write, structure, outline, or review white papers, technical reports, position papers, policy briefs, or evidence-driven arguments. Also use for B2B content marketing strategy, technical documentation requiring authority, academic-style reports, or any document that must educate while persuading. Use when someone asks \"how should I structure this\" for a long-form document.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fwhite-paper-writing\u002FSKILL.md",[],[],{"skillNo":57,"name":58,"nameZh":6,"description":59,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":60,"categories":61,"keywords":62,"publishedAt":19},"01M7AT16BJRJJKK2MASXRZS76G","prose-technical-mechanics","Use when editing or writing prose at a technical\u002Fmechanical level: choosing between Anglo-Saxon and Latinate diction, placing information in topic vs stress positions, writing periodic vs cumulative sentences, handling sentence rhythm, punctuation as logic, transitions, paragraph structure, or phonesthetics. Use when someone wants to understand WHY prose works or doesn't, not just surface-level style advice. Triggers on word choice, sentence structure, emphasis, rhythm, diction, or any deep writing mechanics question.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fprose-technical-mechanics\u002FSKILL.md",[],[],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":19},"01M7AT0W9HT0ZZ9Y287NW99MR3","prose-mechanics","Use when writing, editing, or analyzing prose style; when trying to avoid generic AI-sounding writing; when applying specific prose styles such as journalistic, academic, narrative, KJV biblical, Hemingway plain, Latinate\u002FJamesian, Orwellian, Augustan, Victorian, Early Modern (Bacon, Browne, Milton), or free indirect discourse; or when coaching someone on writing voice and mechanics. Triggers on any request to improve, analyze, or write in a specific prose style, to fix \"AI-sounding\" writing, to make prose more vivid or concrete, to vary sentence rhythm, or to understand how a particular writer achieves their effects.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fprose-mechanics\u002FSKILL.md",[],[],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":19},"01M7AT0KR635JX66X2X6F3E2CP","narrative-structure","Use when structuring stories, pitches, presentations, white papers, or any persuasive narrative; analyzing the hero's journey in fiction or non-fiction; applying StoryBrand for business communication; designing narrative arcs for thought leadership; or understanding Vogler, Snyder, Propp, or Campbell frameworks. Triggers on any narrative structure, storytelling, pitch structure, StoryBrand, hero's journey, three-act structure, Freytag, Save the Cat, narrative arc, or \"how should I structure this story\u002Fpitch\u002Fpresentation\" request.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fnarrative-structure\u002FSKILL.md",[],[],{"skillNo":78,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":83,"publishedAt":19},"01M7AT0AKGW09NKKZDYBT8VRDE","legal-brief-writing","Use when writing legal briefs, motions, or any court filing; when analyzing judicial cognition and persuasion; when applying elite brief-writing techniques from Roberts, Clement, Kagan, Garner; or when structuring legal arguments for maximum persuasion. Also use for autistic\u002Fneurodivergent advocates who need to understand how their natural communication style maps to elite legal advocacy. Triggers on any legal writing, brief drafting, motion preparation, or legal argument structuring request.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Flegal-brief-writing\u002FSKILL.md",[],[],{"skillNo":85,"name":86,"nameZh":6,"description":87,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":88,"categories":89,"keywords":90,"publishedAt":19},"01M7AT00VWPS4Q8B5VEMXX9NQT","journal-publishability-criteria","Use whenever a research paper, article or preprint is being written, revised, reviewed or published, and specifically before every publication of docs\u002Fpaper.md. Covers what legitimate journals actually gate on, in order of gating — the editor's desk screen (scope, novelty, package problems), what reviewers judge (methodological soundness, claims the evidence can carry, statistical and qualitative reporting), EQUATOR reporting guidelines, the ICMJE authorship criteria, COPE's view of text recycling, the near-universal generative-AI policy (no AI author, disclosure, author responsibility), TOP 2025 transparency standards, how to verify a journal is legitimate, submitting without an institution, and a pre-submission checklist. Triggers on requests to evaluate, review, prepare or submit a paper, to judge publishability, to write a declarations or data-availability section, to disclose AI use, or to check a journal.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fjournal-publishability-criteria\u002FSKILL.md",[],[]]