Add per-course content roots and course scaffold

.github/workflows/quality.yml

@@ -0,0 +1,32 @@
+name: Quality
+
+on:
+  pull_request:
+  push:
+    branches:
+      - main
+
+jobs:
+  quality:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+
+      - name: Setup Node
+        uses: actions/setup-node@v6
+        with:
+          node-version: 24
+          cache: npm
+
+      - name: Install dependencies
+        run: npm ci
+
+      - name: Lint
+        run: npm run lint
+
+      - name: Check formatting
+        run: npm run format:check
+
+      - name: Run tests
+        run: npm test

.husky/pre-commit

@@ -0,0 +1,2 @@
+#!/usr/bin/env sh
+npx lint-staged

.prettierignore

@@ -0,0 +1,6 @@
+node_modules/
+docs/.vitepress/dist/
+.git-history-backup-20260519/
+coverage/
+artifacts/
+temp/

README.md

@@ -7,14 +7,19 @@ This repository is no longer only the `Learn Harness Engineering` course. It now
 - a reusable VitePress course platform
 - one active course selected at build time
 - multiple course manifests under `courses/`
+- isolated course content under `courses/<slug>/docs`
 
 ## Current structure
 
 - active course selector: [docs/.vitepress/active-course.mjs](/Users/matheuspuppe/Desktop/estudo/course-builder/docs/.vitepress/active-course.mjs:1)
 - active-course wrapper: [docs/.vitepress/course.manifest.mjs](/Users/matheuspuppe/Desktop/estudo/course-builder/docs/.vitepress/course.manifest.mjs:1)
+- builder shell: [docs/.vitepress/config.mts](/Users/matheuspuppe/Desktop/estudo/course-builder/docs/.vitepress/config.mts:1)
 - course manifests:
   - [courses/learn-harness-engineering/course.manifest.mjs](/Users/matheuspuppe/Desktop/estudo/course-builder/courses/learn-harness-engineering/course.manifest.mjs:1)
   - [courses/agent-ops-bootcamp/course.manifest.mjs](/Users/matheuspuppe/Desktop/estudo/course-builder/courses/agent-ops-bootcamp/course.manifest.mjs:1)
+- course content roots:
+  - [courses/learn-harness-engineering/docs](/Users/matheuspuppe/Desktop/estudo/course-builder/courses/learn-harness-engineering/docs:1)
+  - [courses/agent-ops-bootcamp/docs](/Users/matheuspuppe/Desktop/estudo/course-builder/courses/agent-ops-bootcamp/docs:1)
 
 ## Commands
 
@@ -29,6 +34,7 @@ To inspect and switch courses:
 
 ```bash
 npm run course:list
+npm run course:new -- my-new-course "My New Course"
 npm run course:activate -- learn-harness-engineering
 npm run course:activate -- agent-ops-bootcamp
 ```
@@ -44,22 +50,22 @@ Each course manifest defines:
 - `curriculum`
 - `homeByLocale`
 
-The VitePress layer reads the active manifest and builds nav, sidebar, branding, and homepage from it.
+The VitePress layer reads the active manifest and points `srcDir` at that course's own `docs` directory. Nav, sidebar, branding, homepage, and markdown content all come from the active course.
 
 ## Creating a new course
 
 1. Create `courses/<slug>/course.manifest.mjs`.
-2. Start from one of the existing manifests.
-3. Define the course `site`, `brand`, `theme`, `locales`, `curriculum`, and `homeByLocale`.
-4. Switch to it with `npm run course:activate -- <slug>`.
-5. Update `docs/` markdown content so it matches the new curriculum.
+2. Or scaffold both manifesto and docs with `npm run course:new -- <slug> "Course Title"`.
+3. Create `courses/<slug>/docs/` if you did not use the scaffold command.
+4. Define the course `site`, `brand`, `theme`, `locales`, `curriculum`, and `homeByLocale`.
+5. Add markdown content under `courses/<slug>/docs/<locale>/...` to match the curriculum.
+6. Switch to it with `npm run course:activate -- <slug>`.
 
 ## Current limitations
 
 - only one course is active per build
-- `docs/` content is still mostly from `Learn Harness Engineering`
 - the current manifest format is code-first, not CMS-driven
-- the public docs URLs still depend on each course's configured `site.base`
+- only one course is active per build
 
 ## Validation
 

WHITELABEL.md

@@ -6,6 +6,7 @@ O projeto agora está estruturado para suportar múltiplos cursos com um curso a
 
 - curso ativo: [docs/.vitepress/active-course.mjs](/Users/matheuspuppe/Desktop/estudo/course-builder/docs/.vitepress/active-course.mjs:1)
 - manifesto do curso atual: [courses/learn-harness-engineering/course.manifest.mjs](/Users/matheuspuppe/Desktop/estudo/course-builder/courses/learn-harness-engineering/course.manifest.mjs:1)
+- conteúdo do curso atual: [courses/learn-harness-engineering/docs](/Users/matheuspuppe/Desktop/estudo/course-builder/courses/learn-harness-engineering/docs:1)
 - wrappers compatíveis com a camada antiga:
   - [docs/.vitepress/course.config.mjs](/Users/matheuspuppe/Desktop/estudo/course-builder/docs/.vitepress/course.config.mjs:1)
   - [docs/.vitepress/course-content.mjs](/Users/matheuspuppe/Desktop/estudo/course-builder/docs/.vitepress/course-content.mjs:1)
@@ -19,13 +20,15 @@ O projeto agora está estruturado para suportar múltiplos cursos com um curso a
 - manifesto de locales e labels de UI
 - currículo por locale
 - conteúdo da homepage por locale
+- conteúdo markdown por curso
 - seleção do curso ativo
 
 ## Como criar outro curso
 
 1. Crie uma pasta `courses/<slug>/`.
-2. Copie [courses/learn-harness-engineering/course.manifest.mjs](/Users/matheuspuppe/Desktop/estudo/course-builder/courses/learn-harness-engineering/course.manifest.mjs:1) para o novo slug.
-3. Edite no manifesto:
+2. Crie `courses/<slug>/docs/`.
+3. Copie [courses/learn-harness-engineering/course.manifest.mjs](/Users/matheuspuppe/Desktop/estudo/course-builder/courses/learn-harness-engineering/course.manifest.mjs:1) para o novo slug.
+4. Edite no manifesto:
    - `site.title`
    - `site.description`
    - `site.base`
@@ -35,17 +38,17 @@ O projeto agora está estruturado para suportar múltiplos cursos com um curso a
    - `locales`
    - `curriculum`
    - `homeByLocale`
-4. Aponte [docs/.vitepress/active-course.mjs](/Users/matheuspuppe/Desktop/estudo/course-builder/docs/.vitepress/active-course.mjs:1) para o novo manifesto.
-5. Reescreva o conteúdo real em `docs/<locale>/` para bater com o currículo do novo curso.
+5. Adicione o conteúdo real em `courses/<slug>/docs/<locale>/` para bater com o currículo do novo curso.
+6. Ative o curso.
 
 Ou use os comandos:
 
 - `npm run course:list`
+- `npm run course:new -- <slug> "Course Title"`
 - `npm run course:activate -- <slug>`
 
 ## Limitações atuais
 
-- o conteúdo markdown em `docs/` ainda é do curso `Learn Harness Engineering`
 - o manifesto do curso atual ainda concentra muito conteúdo inline
 - ainda existe apenas um curso ativo por build
-- `README.md` ainda descreve o curso atual, não a plataforma
+- o builder ainda depende de manifests em código, não de uma camada editorial

courses/agent-ops-bootcamp/course.manifest.mjs

@@ -4,33 +4,30 @@ const courseManifest = {
     description:
       "A compact course on operating, debugging, and shipping agentic workflows in production teams.",
     base: "/course-builder/",
-    repoTreeUrl: "https://github.com/puppe1990/course-builder/tree/main"
+    repoTreeUrl: "https://github.com/puppe1990/course-builder/tree/main",
   },
   brand: {
-    logo:
-      'data:image/svg+xml;utf8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="%230B6E4F" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"><path d="M4 12h16"/><path d="M12 4v16"/><circle cx="12" cy="12" r="8"/></svg>'
+    logo: 'data:image/svg+xml;utf8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="%230B6E4F" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"><path d="M4 12h16"/><path d="M12 4v16"/><circle cx="12" cy="12" r="8"/></svg>',
   },
   theme: {
     fontStylesheets: [
-      "https://fonts.googleapis.com/css2?family=Newsreader:opsz,wght@6..72,400;6..72,600&family=Plus+Jakarta+Sans:wght@400;500;600;700&family=JetBrains+Mono:wght@400;500&display=swap"
+      "https://fonts.googleapis.com/css2?family=Newsreader:opsz,wght@6..72,400;6..72,600&family=Plus+Jakarta+Sans:wght@400;500;600;700&family=JetBrains+Mono:wght@400;500&display=swap",
     ],
     typography: {
-      body:
-        "'Plus Jakarta Sans', ui-sans-serif, system-ui, -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif",
+      body: "'Plus Jakarta Sans', ui-sans-serif, system-ui, -apple-system, BlinkMacSystemFont, 'Segoe UI', sans-serif",
       heading:
         "'Newsreader', ui-serif, Georgia, Cambria, 'Times New Roman', Times, serif",
-      mono:
-        "'JetBrains Mono', 'Fira Code', ui-monospace, SFMono-Regular, Menlo, Consolas, monospace",
-      mermaid: "'Plus Jakarta Sans', sans-serif"
+      mono: "'JetBrains Mono', 'Fira Code', ui-monospace, SFMono-Regular, Menlo, Consolas, monospace",
+      mermaid: "'Plus Jakarta Sans', sans-serif",
     },
     layout: {
       maxWidth: "1280px",
       sidebarWidth: "280px",
-      cardRadius: "14px"
+      cardRadius: "14px",
     },
     effects: {
       featureShadow: "0 18px 48px rgba(11, 110, 79, 0.14)",
-      featureShadowDark: "0 18px 48px rgba(0, 0, 0, 0.4)"
+      featureShadowDark: "0 18px 48px rgba(0, 0, 0, 0.4)",
     },
     colors: {
       light: {
@@ -47,7 +44,7 @@ const courseManifest = {
         brandSoft: "rgba(11, 110, 79, 0.12)",
         divider: "rgba(8, 32, 20, 0.1)",
         sidebarBg: "#ECF5F0",
-        navBg: "#F6FBF8"
+        navBg: "#F6FBF8",
       },
       dark: {
         bg: "#0E1712",
@@ -63,9 +60,9 @@ const courseManifest = {
         brandSoft: "rgba(57, 163, 127, 0.16)",
         divider: "rgba(255, 255, 255, 0.12)",
         sidebarBg: "#0A120D",
-        navBg: "#0E1712"
-      }
-    }
+        navBg: "#0E1712",
+      },
+    },
   },
   locales: [
     {
@@ -78,71 +75,80 @@ const courseManifest = {
         resources: "Toolkit",
         skills: "Playbooks",
         resourceLibrary: "Toolkit",
-        tryHarness: "Open Templates ↗"
-      }
-    }
+        tryHarness: "Open Templates ↗",
+      },
+    },
   ],
   curriculum: {
     en: {
       lectures: [
         { text: "Welcome", link: "/en/" },
-        { text: "Operate Agents with Guardrails", link: "/en/lectures/operate-agents-with-guardrails/" },
-        { text: "Debug Long-Running Workflows", link: "/en/lectures/debug-long-running-workflows/" },
-        { text: "Ship with Runtime Feedback", link: "/en/lectures/ship-with-runtime-feedback/" }
+        {
+          text: "Operate Agents with Guardrails",
+          link: "/en/lectures/operate-agents-with-guardrails/",
+        },
+        {
+          text: "Debug Long-Running Workflows",
+          link: "/en/lectures/debug-long-running-workflows/",
+        },
+        {
+          text: "Ship with Runtime Feedback",
+          link: "/en/lectures/ship-with-runtime-feedback/",
+        },
       ],
       projects: [
         { text: "Overview", link: "/en/projects/" },
         { text: "Incident Drill", link: "/en/projects/incident-drill/" },
-        { text: "Release Checklist", link: "/en/projects/release-checklist/" }
+        { text: "Release Checklist", link: "/en/projects/release-checklist/" },
       ],
       resources: [
         { text: "Overview", link: "/en/resources/" },
         { text: "Runbooks", link: "/en/resources/templates/" },
         { text: "Reference", link: "/en/resources/reference/" },
-        { text: "Advanced Ops", link: "/en/resources/openai-advanced/" }
+        { text: "Advanced Ops", link: "/en/resources/openai-advanced/" },
       ],
-      skills: [{ text: "Playbooks", link: "/en/skills/" }]
-    }
+      skills: [{ text: "Playbooks", link: "/en/skills/" }],
+    },
   },
   homeByLocale: {
     en: {
       hero: {
         title: "Welcome to Agent Ops Bootcamp",
         intro: [
           "Agent Ops Bootcamp focuses on how teams run, observe, and correct agentic systems under production constraints.",
-          "The course is intentionally compact: a few modules, a few labs, and a toolkit you can adapt to your own operating model."
+          "The course is intentionally compact: a few modules, a few labs, and a toolkit you can adapt to your own operating model.",
         ],
         references: [
           {
             label: "Operational feedback loops",
-            href: "https://openai.com/index/harness-engineering/"
-          }
-        ]
+            href: "https://openai.com/index/harness-engineering/",
+          },
+        ],
       },
       sections: {
         start: "Start here",
         mechanism: "Operating loop",
         learn: "You will practice",
-        next: "Next moves"
+        next: "Next moves",
       },
       startText:
         "This sample course exists to prove the platform can host a second curriculum with different naming, visual identity, and positioning.",
       cards: [
         {
           title: "Modules",
           body: "Short theory units about operating agents safely.",
-          href: "./lectures/operate-agents-with-guardrails/"
+          href: "./lectures/operate-agents-with-guardrails/",
         },
         {
           title: "Labs",
           body: "Hands-on operational drills.",
-          href: "./projects/"
+          href: "./projects/",
         },
         {
           title: "Toolkit",
           body: "Runbooks and reusable templates.",
-          href: "./resources/"
-        }
+          href: "./resources/",
+        },
       ],
       mechanismText:
         "Production agent work improves when runtime feedback and verification are part of the operating loop.",
@@ -155,23 +161,23 @@ const courseManifest = {
       learnItems: [
         "Define operator-facing guardrails.",
         "Treat logs and runtime feedback as first-class inputs.",
-        "Run lightweight drills before real incidents."
+        "Run lightweight drills before real incidents.",
       ],
       nextIntro: "The sample path is intentionally small:",
       nextSteps: [
         {
           title: "Module 1",
           description: "Start with guardrails.",
-          href: "./lectures/operate-agents-with-guardrails/"
+          href: "./lectures/operate-agents-with-guardrails/",
         },
         {
           title: "Lab 1",
           description: "Run an incident drill.",
-          href: "./projects/incident-drill/"
-        }
-      ]
-    }
-  }
+          href: "./projects/incident-drill/",
+        },
+      ],
+    },
+  },
 };
 
 export default courseManifest;

courses/agent-ops-bootcamp/docs/en/lectures/debug-long-running-workflows/index.md

@@ -0,0 +1,14 @@
+# Debug Long-Running Workflows
+
+Long-running agents rarely fail in one obvious place. They drift, loop, or lose context across steps.
+
+## What to inspect
+
+- Execution logs
+- State transitions
+- Retry behavior
+- Handoff boundaries
+
+## Debugging principle
+
+Treat every long-running workflow as an observable system, not a black box prompt.

courses/agent-ops-bootcamp/docs/en/lectures/operate-agents-with-guardrails/index.md

@@ -0,0 +1,13 @@
+# Operate Agents with Guardrails
+
+Agent systems need explicit operating boundaries. This module shows how teams define ownership, stop conditions, and escalation paths before runtime issues become incidents.
+
+## Core rules
+
+- Define who can trigger the workflow.
+- Define what the agent can change without approval.
+- Define when the system must pause and ask for human review.
+
+## Outcome
+
+By the end of this module, you should have a lightweight guardrail policy that can be attached to any agent workflow.

courses/agent-ops-bootcamp/docs/en/lectures/ship-with-runtime-feedback/index.md

@@ -0,0 +1,14 @@
+# Ship with Runtime Feedback
+
+Shipping agent systems safely requires a feedback loop between runtime behavior and product decisions.
+
+## Runtime inputs
+
+- Error classes
+- Human overrides
+- Repeated retries
+- Failed postconditions
+
+## Practice
+
+Turn the most common runtime failures into explicit operating rules and release checks.

courses/agent-ops-bootcamp/docs/en/projects/incident-drill/index.md

@@ -0,0 +1,7 @@
+# Incident Drill
+
+Simulate a production issue where an agent keeps retrying a task with stale context.
+
+## Deliverable
+
+Write the operator checklist your team should follow in the first ten minutes of the incident.

courses/agent-ops-bootcamp/docs/en/projects/index.md

@@ -0,0 +1,3 @@
+# Labs
+
+Use these labs to pressure-test your operating model before a real release.