feat(plugin,website): split elaborate into orchestrator + elaborator agent, add changelog page, update docs

Plugin changes:
- Create elaborator agent (plugin/agents/elaborator.md) running on Opus with fresh context
- Create elaboration-start internal skill with Phases 2-7 (context: fork, agent: ai-dlc:elaborator)
- Slim elaborate skill to Phase 0 + pre-compaction orchestrator (1567 -> 267 lines)

Website changes:
- Add /changelog/ page rendering CHANGELOG.md with color-coded badges and linkable anchors
- Rewrite hats.md: expand from "The Four Hats" (4) to "The Hat System" (14 hats across workflows)
- Fix workflows.md: correct hat sequences, add Design workflow, add per-unit workflows section
- Update all "Four Hats" references across homepage, about, big-picture, start-here, nav, breadcrumbs

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: jwaldrip

plugin/agents/elaborator.md

@@ -0,0 +1,89 @@
+---
+model: opus
+allowedTools:
+  - Read
+  - Write
+  - Glob
+  - Grep
+  - Bash
+  - Agent
+  - Skill
+  - WebSearch
+  - WebFetch
+  - AskUserQuestion
+  - ListMcpResourcesTool
+  - ReadMcpResourceTool
+  # MCP read-only tool patterns
+  - "mcp__*__read*"
+  - "mcp__*__get*"
+  - "mcp__*__list*"
+  - "mcp__*__search*"
+  - "mcp__*__query*"
+  - "mcp__*__ask*"
+  - "mcp__*__resolve*"
+  - "mcp__*__fetch*"
+  - "mcp__*__lookup*"
+  - "mcp__*__analyze*"
+  - "mcp__*__describe*"
+  - "mcp__*__explain*"
+  - "mcp__*__memory"
+  # Ticketing provider write tools (epic/ticket creation during elaboration)
+  - "mcp__*__create*issue*"
+  - "mcp__*__create*ticket*"
+  - "mcp__*__create*epic*"
+  - "mcp__*__update*issue*"
+  - "mcp__*__update*ticket*"
+  - "mcp__*__add*comment*"
+disallowedTools:
+  - Edit
+  - NotebookEdit
+  - EnterPlanMode
+  - ExitPlanMode
+  - EnterWorktree
+  - TeamCreate
+  - TeamDelete
+  - SendMessage
+  - TaskCreate
+  - TaskUpdate
+  - TaskGet
+  - TaskList
+---
+
+# AI-DLC Elaborator
+
+You are the **AI-DLC Elaborator** — a specialized agent that runs the full elaboration workflow from a pre-compacted brief. You operate with a fresh context window on Opus to maximize the quality of domain discovery and spec writing.
+
+## Your Mission
+
+You receive an **Elaboration Brief** containing:
+- The user's intent (what they want to build and why)
+- Requirements and constraints gathered from conversation
+- Technical context (systems, APIs, codebases)
+- Project metadata (root path, cowork status, workspace)
+- Any existing context from provider discovery
+
+Your job is to take this brief and execute the full elaboration protocol:
+1. **Clarify Requirements** — Ask targeted follow-up questions based on the brief
+2. **Create Intent Worktree** — Set up the branch and discovery log
+3. **Domain Discovery** — Deep technical exploration using Explore subagents, MCP tools, and web research
+4. **Discover Hats & Select Workflow** — Present workflow options
+5. **Define Success Criteria** — Collaborate on verifiable criteria
+6. **Decompose into Units** — Break work into independent, well-specified units
+7. **Write Artifacts** — Produce intent.md, unit files, wireframes, and tickets
+8. **Handoff** — Present summary and next steps
+
+## Key Principles
+
+- **Depth over speed**: Elaboration is not a quick phase. Build deep domain understanding.
+- **Verify, don't guess**: Use Explore subagents to read real code, query real APIs, and map real schemas.
+- **Collaborate**: Use AskUserQuestion to validate your understanding at each major step.
+- **Persist findings**: Write to discovery.md as you go — your context window is precious.
+- **Specificity**: Every unit must be detailed enough that a builder with zero prior context builds the right thing.
+
+## Subagent Usage
+
+You can and should spawn subagents for parallel research:
+- Use `Explore` subagent type for deep codebase and API exploration
+- Use `general-purpose` subagent type for design file analysis (needs MCP tool access)
+- Launch multiple subagents in parallel for independent research paths
+- Summarize findings to the user as subagents return results

plugin/skills/elaborate/SKILL.md

@@ -38,15 +38,10 @@ export default function AboutPage() {
 						each phase of development gets proper attention.
 					</p>
 
-					<h2>The Four Hats</h2>
+					<h2>The Hat System</h2>
 
-					<h3>Researcher</h3>
 					<p>
-						The Researcher gathers context and understanding before any
-						implementation begins. This phase involves reading existing code,
-						understanding requirements, researching patterns, and identifying
-						constraints. The goal is to build a complete mental model of the
-						problem space.
+						AI-DLC organizes work through hats — distinct mindsets that keep each phase of development focused. The default construction workflow uses three core hats, while specialized workflows add hats for security testing, design, TDD, and scientific debugging.
 					</p>
 
 					<h3>Planner</h3>
@@ -73,6 +68,10 @@ export default function AboutPage() {
 						is ready for production.
 					</p>
 
+					<p>
+						Beyond the core three, AI-DLC includes specialized hats like Designer for UI/UX work, Red Team and Blue Team for adversarial security testing, and Observer, Hypothesizer, Experimenter, and Analyst for scientific debugging. See the <a href="/docs/hats/">full hat reference</a> for details.
+					</p>
+
 					<h2>Units of Work</h2>
 					<p>
 						Work is organized into units. Each unit is a focused piece of

plugin/skills/elaboration-start/SKILL.md

@@ -40,8 +40,8 @@ export default function BigPicturePage() {
 						color="blue"
 					/>
 					<ExplanationCard
-						title="The Four Hats"
-						description="Each hat represents a distinct mindset: Researcher gathers context, Planner designs the approach, Builder executes, and Reviewer validates quality."
+						title="The Hat System"
+						description="Each hat represents a distinct mindset. The core workflow uses Planner, Builder, and Reviewer, with specialized hats for security testing, design, debugging, and TDD."
 						href="/docs/hats/"
 						color="purple"
 					/>

website/app/about/page.tsx

@@ -0,0 +1,116 @@
+import { getChangelog } from "@/lib/changelog"
+import type { Metadata } from "next"
+
+export const metadata: Metadata = {
+	title: "Changelog",
+	description:
+		"What's new in AI-DLC — a complete history of features, fixes, and changes.",
+	openGraph: {
+		title: "AI-DLC Changelog",
+		description:
+			"What's new in AI-DLC — a complete history of features, fixes, and changes.",
+	},
+}
+
+function formatDate(dateString: string): string {
+	const date = new Date(`${dateString}T00:00:00`)
+	return date.toLocaleDateString("en-US", {
+		year: "numeric",
+		month: "long",
+		day: "numeric",
+	})
+}
+
+function sectionTypeColor(type: string): string {
+	switch (type.toLowerCase()) {
+		case "added":
+			return "bg-green-100 text-green-800 dark:bg-green-900/30 dark:text-green-400"
+		case "fixed":
+			return "bg-blue-100 text-blue-800 dark:bg-blue-900/30 dark:text-blue-400"
+		case "changed":
+			return "bg-yellow-100 text-yellow-800 dark:bg-yellow-900/30 dark:text-yellow-400"
+		default:
+			return "bg-gray-100 text-gray-800 dark:bg-gray-800 dark:text-gray-400"
+	}
+}
+
+export default function ChangelogPage() {
+	const entries = getChangelog()
+
+	return (
+		<div className="mx-auto max-w-3xl px-4 py-8 lg:py-12">
+			<header className="mb-12">
+				<h1 className="text-4xl font-bold tracking-tight sm:text-5xl">
+					Changelog
+				</h1>
+				<p className="mt-2 text-lg text-gray-600 dark:text-gray-400">
+					A complete history of features, fixes, and changes in AI-DLC.
+				</p>
+			</header>
+
+			{entries.length === 0 ? (
+				<div className="rounded-lg border border-gray-200 bg-gray-50 p-8 text-center dark:border-gray-800 dark:bg-gray-900">
+					<p className="text-gray-600 dark:text-gray-400">
+						No changelog entries yet.
+					</p>
+				</div>
+			) : (
+				<div className="space-y-12">
+					{entries.map((entry) => (
+						<section
+							key={entry.version}
+							id={`v${entry.version}`}
+							className="scroll-mt-24"
+						>
+							<div className="mb-4 flex flex-wrap items-baseline gap-3">
+								<a
+									href={`#v${entry.version}`}
+									className="text-2xl font-bold tracking-tight hover:text-blue-600 dark:hover:text-blue-400"
+								>
+									v{entry.version}
+								</a>
+								<time className="text-sm text-gray-500 dark:text-gray-500">
+									{formatDate(entry.date)}
+								</time>
+							</div>
+
+							{entry.sections.length > 0 ? (
+								<div className="space-y-4">
+									{entry.sections.map((section) => (
+										<div key={section.type}>
+											<span
+												className={`inline-block rounded-full px-2.5 py-0.5 text-xs font-medium ${sectionTypeColor(section.type)}`}
+											>
+												{section.type}
+											</span>
+											<ul className="mt-2 space-y-1">
+												{section.items.map((item, index) => (
+													<li
+														key={index}
+														className="flex gap-2 text-gray-700 dark:text-gray-300"
+													>
+														<span className="mt-1.5 h-1.5 w-1.5 flex-shrink-0 rounded-full bg-gray-400 dark:bg-gray-600" />
+														<span>{item}</span>
+													</li>
+												))}
+											</ul>
+										</div>
+									))}
+								</div>
+							) : (
+								<p className="text-sm text-gray-500 dark:text-gray-500">
+									No notable changes.
+								</p>
+							)}
+
+							{entry !==
+								entries[entries.length - 1] && (
+								<div className="mt-8 border-b border-gray-200 dark:border-gray-800" />
+							)}
+						</section>
+					))}
+				</div>
+			)}
+		</div>
+	)
+}

website/app/big-picture/page.tsx

@@ -23,7 +23,7 @@ const segmentLabels: Record<string, string> = {
 	"quick-start": "Quick Start",
 	installation: "Installation",
 	concepts: "Concepts",
-	hats: "The Four Hats",
+	hats: "Hats",
 	community: "Community",
 	"example-feature": "Feature Example",
 	"example-bugfix": "Bugfix Example",

website/app/changelog/page.tsx

@@ -2,30 +2,6 @@ import { getAllBlogPosts } from "@/lib/blog"
 import Link from "next/link"
 
 const hats = [
-	{
-		name: "Researcher",
-		icon: (
-			<svg
-				className="h-8 w-8"
-				fill="none"
-				viewBox="0 0 24 24"
-				stroke="currentColor"
-				aria-hidden="true"
-			>
-				<path
-					strokeLinecap="round"
-					strokeLinejoin="round"
-					strokeWidth={1.5}
-					d="M21 21l-6-6m2-5a7 7 0 11-14 0 7 7 0 0114 0z"
-				/>
-			</svg>
-		),
-		description:
-			"Gather context, understand requirements, and research patterns.",
-		color: "text-blue-600 dark:text-blue-400",
-		bgColor: "bg-blue-50 dark:bg-blue-950/50",
-		borderColor: "border-blue-200 dark:border-blue-900",
-	},
 	{
 		name: "Planner",
 		icon: (
@@ -456,13 +432,14 @@ export default function Home() {
 			<section className="bg-gray-50 px-4 py-16 dark:bg-gray-900/50">
 				<div className="mx-auto max-w-6xl">
 					<div className="mb-12 text-center">
-						<h2 className="mb-4 text-3xl font-bold">Four Hats, One Workflow</h2>
+						<h2 className="mb-4 text-3xl font-bold">Hats Drive Every Phase</h2>
 						<p className="mx-auto max-w-2xl text-lg text-gray-600 dark:text-gray-400">
-							AI-DLC uses a hat-based approach to separate concerns. Each hat
-							represents a distinct mindset and set of responsibilities.
+							AI-DLC uses a hat-based approach to separate concerns. The default
+							workflow uses three core hats, with specialized hats available for
+							security, design, testing, and debugging workflows.
 						</p>
 					</div>
-					<div className="grid gap-4 sm:grid-cols-2 lg:grid-cols-4">
+					<div className="grid gap-4 sm:grid-cols-2 lg:grid-cols-3">
 						{hats.map((hat) => (
 							<div
 								key={hat.name}

website/app/components/navigation/Breadcrumbs.tsx

@@ -34,7 +34,7 @@ const steps: Step[] = [
 			},
 			{
 				id: "hats",
-				label: "Learn about the Four Hats",
+				label: "Learn about the Hat System",
 				href: "/docs/hats/",
 			},
 			{