{"id":"20bec32c-9114-42b0-8e20-8caaadda9b84","entityType":"agent","slug":"clawhub-ericn26-star-eric-knowledge-digest","name":"Knowledge Digest","canonicalUrl":"https://www.xpersona.co/agent/clawhub-ericn26-star-eric-knowledge-digest","canonicalPath":"/agent/clawhub-ericn26-star-eric-knowledge-digest","generatedAt":"2026-10-13T00:33:30.275Z","source":"CLAWHUB","claimStatus":"UNCLAIMED","verificationTier":"NONE","summary":{"evidence":{"source":"editorial-content","verified":true,"confidence":"high","updatedAt":"2026-10-11T23:43:24.843Z","emptyReason":null},"description":"Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, an... Skill: Knowledge Digest Owner: ericn26-star Summary: Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, an... Tags: latest:11.0.0 Version history: v1.2.0 | 2026-05-05T10:17:41.510Z | auto Version 1.2.0 — No changes detected in this release. - No file changes or updates were made compared to the previous version. v1","descriptionLabel":"Technical summary","evidenceSummary":"Capability contract not published. No trust telemetry is available yet. 974 downloads reported by the source. Last updated 10/11/2026.","installCommand":"clawhub skill install s17a4j37w3f2wyke5hajt5vtad85wyxf:eric-knowledge-digest","sourceUrl":"https://clawhub.ai/ericn26-star/eric-knowledge-digest","homepage":"https://clawhub.ai/ericn26-star/skills/eric-knowledge-digest","primaryLinks":[{"label":"View on ClawHub","url":"https://clawhub.ai/ericn26-star/eric-knowledge-digest","kind":"source"},{"label":"Homepage","url":"https://clawhub.ai/ericn26-star/skills/eric-knowledge-digest","kind":"homepage"}],"safetyScore":84,"overallRank":62,"popularityScore":60,"trustScore":null,"claimedByName":null,"isOwner":false,"seoDescription":"Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, an..."},"coverage":{"evidence":{"source":"public-profile","verified":false,"confidence":"medium","updatedAt":"2026-10-11T23:43:24.843Z","emptyReason":null},"protocols":[{"protocol":"OPENCLEW","label":"OpenClaw","status":"self-declared","notes":"Declared in the public agent profile."}],"capabilities":[],"verifiedCount":0,"selfDeclaredCount":1,"capabilityMatrix":{"rows":[{"key":"OPENCLEW","type":"protocol","support":"unknown","confidenceSource":"profile","notes":"Listed on profile"}],"flattenedTokens":"protocol:OPENCLEW|unknown|profile"}},"adoption":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T23:43:24.843Z","emptyReason":null},"stars":null,"forks":null,"downloads":974,"likes":null,"task":null,"library":null,"packageName":null,"latestVersion":"1.2.0","tractionLabel":"974 downloads"},"release":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T23:43:24.828Z","emptyReason":null},"lastUpdatedAt":"2026-10-11T23:43:24.843Z","lastCrawledAt":"2026-10-11T23:43:24.828Z","lastIndexedAt":null,"nextCrawlAt":"2026-10-12T23:43:24.828Z","lastVerifiedAt":null,"highlights":[{"version":"1.2.0","createdAt":"2026-05-05T10:17:41.510Z","changelog":"Version 1.2.0 — No changes detected in this release. - No file changes or updates were made compared to the previous version.","fileCount":2,"zipByteSize":10473},{"version":"1.0.5","createdAt":"2026-05-05T03:56:57.393Z","changelog":"No changes detected in this version. - Version 1.0.5 is identical to the previous release; no file updates or feature modifications found.","fileCount":2,"zipByteSize":10471},{"version":"1.0.10","createdAt":"2026-05-04T17:56:35.105Z","changelog":"No file changes detected in this version. No updates or new features were introduced.","fileCount":2,"zipByteSize":10473},{"version":"11.0.0","createdAt":"2026-05-03T10:01:24.635Z","changelog":"No user-visible changes in this release (version 11.0.0). No file or documentation modifications were detected.","fileCount":3,"zipByteSize":11593},{"version":"10.0.0","createdAt":"2026-05-03T05:02:35.298Z","changelog":"KnowledgeDigest 10.0.0 — Major Upgrade: Unified conversion of textbooks/PDFs into multimodal, personalized learning assets. - Converts educational materials or topics into personalized notes, quizzes, slides, mind maps, and audio courses, tailored to user’s grade and interests. - Automatically analyzes and structures content, extracting chapters, key concepts, learning objectives, and knowledge dependencies. - Supports multiple output formats: handwritten-style notes, interactive quizzes, slides (PDF/PPTX), mind maps (image/Mermaid), and teacher-student audio courses (MP3). - New workflow for handling cases with or without uploaded materials, including content search and user confirmation steps if no files are provided. - Grade level and interest adaptation for all materials, ensuring suitable language, detail, and style for each user. - Standardized, private delivery: file paths only, no inline previews or autoplay of media.","fileCount":2,"zipByteSize":10472},{"version":"3.0.1","createdAt":"2026-05-02T05:59:20.386Z","changelog":"- No changes detected in this version. - The SKILL.md content remains unchanged from the previous version.","fileCount":2,"zipByteSize":10472},{"version":"3.0.0","createdAt":"2026-05-01T21:56:54.934Z","changelog":"No changes detected in this version. - Version 3.0.0 contains no file changes compared to the previous release. - All workflow logic, output formats, and personalization features remain unchanged.","fileCount":2,"zipByteSize":10473},{"version":"2.0.0","createdAt":"2026-05-01T15:57:06.423Z","changelog":"Version 2.0.0 - No file changes detected; documentation and workflow remain unchanged. - Core features and supported output formats are consistent with previous version. - No new features, bug fixes, or breaking changes introduced in this release.","fileCount":2,"zipByteSize":10472}]},"execution":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No published capability contract is available yet."},"installCommand":"clawhub skill install s17a4j37w3f2wyke5hajt5vtad85wyxf:eric-knowledge-digest","setupComplexity":"low","setupSteps":["Setup complexity is classified as HIGH. You must provision dedicated cloud infrastructure or an isolated VM. Do not run this directly on your local workstation.","Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data."],"contract":{"contractStatus":"missing","authModes":[],"requires":[],"forbidden":[],"supportsMcp":false,"supportsA2a":false,"supportsStreaming":false,"inputSchemaRef":null,"outputSchemaRef":null,"dataRegion":null,"contractUpdatedAt":null,"sourceUpdatedAt":null,"freshnessSeconds":null},"invocationGuide":{"preferredApi":{"snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-ericn26-star-eric-knowledge-digest/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-ericn26-star-eric-knowledge-digest/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-ericn26-star-eric-knowledge-digest/trust"},"curlExamples":["curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-ericn26-star-eric-knowledge-digest/snapshot\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-ericn26-star-eric-knowledge-digest/contract\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-ericn26-star-eric-knowledge-digest/trust\""],"jsonRequestTemplate":{"query":"summarize this repo","constraints":{"maxLatencyMs":2000,"protocolPreference":["OPENCLEW"]}},"jsonResponseTemplate":{"ok":true,"result":{"summary":"...","confidence":0.9},"meta":{"source":"CLAWHUB","generatedAt":"2026-10-13T00:33:30.273Z"}},"retryPolicy":{"maxAttempts":3,"backoffMs":[500,1500,3500],"retryableConditions":["HTTP_429","HTTP_503","NETWORK_TIMEOUT"]}},"endpoints":{"dossierUrl":"https://www.xpersona.co/api/v1/agents/clawhub-ericn26-star-eric-knowledge-digest/dossier","snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-ericn26-star-eric-knowledge-digest/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-ericn26-star-eric-knowledge-digest/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-ericn26-star-eric-knowledge-digest/trust"}},"reliability":{"evidence":{"source":"runtime-metrics","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No trust, reliability, or runtime telemetry is available."},"trust":{"status":"unavailable","handshakeStatus":"UNKNOWN","verificationFreshnessHours":null,"reputationScore":null,"p95LatencyMs":null,"successRate30d":null,"fallbackRate":null,"attempts30d":null,"trustUpdatedAt":null,"trustConfidence":"unknown","sourceUpdatedAt":null,"freshnessSeconds":null},"decisionGuardrails":{"doNotUseIf":["Contract metadata is missing or unavailable for deterministic execution."],"safeUseWhen":[],"riskFlags":["missing_or_unavailable_contract","trust_data_unavailable","schema_references_missing"],"operationalConfidence":"low"},"executionMetrics":{"observedLatencyMsP50":null,"observedLatencyMsP95":null,"estimatedCostUsd":null,"uptime30d":null,"rateLimitRpm":null,"rateLimitBurst":null,"lastVerifiedAt":null,"verificationSource":null},"runtimeMetrics":{"successRate":null,"avgLatencyMs":null,"avgCostUsd":null,"hallucinationRate":null,"retryRate":null,"disputeRate":null,"p50Latency":null,"p95Latency":null,"lastUpdated":null}},"benchmarks":{"evidence":{"source":"no-benchmark-data","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No benchmark suites or observed failure patterns are available."},"suites":[],"failurePatterns":[]},"artifacts":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"high","updatedAt":"2026-10-11T23:43:24.843Z","emptyReason":null},"readme":"Skill: Knowledge Digest\n\nOwner: ericn26-star\n\nSummary: Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, an...\n\nTags: latest:11.0.0\n\nVersion history:\n\nv1.2.0 | 2026-05-05T10:17:41.510Z | auto\n\nVersion 1.2.0 — No changes detected in this release.\n- No file changes or updates were made compared to the previous version.\n\nv1.0.5 | 2026-05-05T03:56:57.393Z | auto\n\nNo changes detected in this version.\n\n- Version 1.0.5 is identical to the previous release; no file updates or feature modifications found.\n\nv1.0.10 | 2026-05-04T17:56:35.105Z | auto\n\nNo file changes detected in this version. No updates or new features were introduced.\n\nv11.0.0 | 2026-05-03T10:01:24.635Z | auto\n\nNo user-visible changes in this release (version 11.0.0).  \nNo file or documentation modifications were detected.\n\nv10.0.0 | 2026-05-03T05:02:35.298Z | auto\n\nKnowledgeDigest 10.0.0 — Major Upgrade: Unified conversion of textbooks/PDFs into multimodal, personalized learning assets.\n\n- Converts educational materials or topics into personalized notes, quizzes, slides, mind maps, and audio courses, tailored to user’s grade and interests.\n- Automatically analyzes and structures content, extracting chapters, key concepts, learning objectives, and knowledge dependencies.\n- Supports multiple output formats: handwritten-style notes, interactive quizzes, slides (PDF/PPTX), mind maps (image/Mermaid), and teacher-student audio courses (MP3).\n- New workflow for handling cases with or without uploaded materials, including content search and user confirmation steps if no files are provided.\n- Grade level and interest adaptation for all materials, ensuring suitable language, detail, and style for each user.\n- Standardized, private delivery: file paths only, no inline previews or autoplay of media.\n\nv3.0.1 | 2026-05-02T05:59:20.386Z | auto\n\n- No changes detected in this version.  \n- The SKILL.md content remains unchanged from the previous version.\n\nv3.0.0 | 2026-05-01T21:56:54.934Z | auto\n\nNo changes detected in this version.  \n- Version 3.0.0 contains no file changes compared to the previous release.  \n- All workflow logic, output formats, and personalization features remain unchanged.\n\nv2.0.0 | 2026-05-01T15:57:06.423Z | auto\n\nVersion 2.0.0\n\n- No file changes detected; documentation and workflow remain unchanged.\n- Core features and supported output formats are consistent with previous version.\n- No new features, bug fixes, or breaking changes introduced in this release.\n\nv1.0.2 | 2026-05-01T12:56:18.502Z | auto\n\nNo user-facing or internal feature changes in this version.\n\n- Version bump only; no code or documentation changes detected.\n- All workflows and features remain unchanged.\n\nv1.0.1 | 2026-05-01T08:56:22.248Z | auto\n\nVersion 1.0.1\n\n- No file or documentation changes were detected in this release.\n- Functionality, workflow, and supported features remain unchanged.\n\nv1.1.0 | 2026-05-01T07:55:26.613Z | auto\n\nNo visible changes detected in this version.\n\n- Functionality, description, and workflow remain unchanged.\n- No modifications found in supported formats or personalization guidelines.\n- Documentation content is identical to previous version.\n\nv1.0.0 | 2026-05-01T06:55:23.711Z | auto\n\nKnowledgeDigest 1.0.0 — Unified Learning Content Converter\n\n- Converts textbooks or PDFs into personalized, multimodal learning materials (handwritten notes, quizzes, slides, audio courses, mind maps).\n- Automatically adapts content for different grade levels and user interests.\n- Supports multiple output formats: annotated notes (PDF), quiz webpages (HTML), slides (PDF/PPTX), mind maps (image/Mermaid), audio courses (MP3).\n- Can generate courseware from user uploads or via search if no material is provided.\n- Extracts structured knowledge, including chapter hierarchy, key concepts, objectives, and builds a knowledge graph for the material.\n- Assets are delivered as file paths only—no inline previews or auto-play for audio/video.\n\nArchive index:\n\nArchive v1.2.0: 2 files, 10473 bytes\n\nFiles: _meta.json (140b), SKILL.md (26234b)\n\nFile v1.2.0:SKILL.md\n\n---\nname: knowledge-digest\ndescription: \"Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, and mind maps. Trigger: learning materials, convert textbook, study notes, quiz generation, slides from PDF, mind map, audio course.\"\n---\n\n# KnowledgeDigest — Unified Learning Content Converter\n\n## Overview\n\nKnowledgeDigest converts textbooks, PDFs, or topic descriptions into personalized, multimodal learning experiences. It analyzes source content, then generates any combination of: handwritten-style notes (PDF), interactive quiz webpages (HTML), slides (PDF+PPTX), mind maps (image+Mermaid), and audio courses (MP3). All output is adapted to the learner's grade level and interests.\n\n## Workflow\n\n### Phase 1: Gather User Input\n\n1. Identify what the user has provided:\n   - Uploaded PDF/textbook file (optional)\n   - Topic/direction description\n   - Grade level (elementary / middle school / high school / university / professional)\n   - Expected output format(s)\n\n2. If no PDF/textbook uploaded and no source materials specified (only topic/direction provided):\n   - Ask user:\n     - **Option A:** \"I have materials, uploading now\"\n     - **Option B:** \"No materials, please search and generate courseware about [topic]\"\n   - If user selects B:\n     - Use search tools to collect authoritative materials on the topic\n     - Organize into structured content, generate a basic courseware PDF\n     - Send PDF to user for confirmation: \"This is the basic material I compiled for [topic], please confirm if usable?\"\n     - Continue after user confirmation\n\n3. **Default output formats** (if user does not specify): mindmap + slides (PDF only) + quiz\n\n### Phase 2: Content Analysis\n\nParse the PDF or structured content to extract:\n\n**Document Parsing:**\n- Identify chapter structure (chapters, sections, subsections)\n- Extract heading hierarchy and table of contents\n- Identify body text, images, tables, formulas, and other elements\n\n**Core Concept Extraction:**\n- Identify core concepts and key terms in each chapter\n- Extract definitions, theorems, formulas, and important content\n- Mark difficult points and key knowledge\n\n**Learning Objective Analysis:**\n- Infer learning objectives for each chapter\n- Identify prerequisite knowledge requirements\n- Analyze dependencies between knowledge points\n\n**Output structured analysis results in this format:**\n\n```json\n{\n  \"document_info\": {\n    \"title\": \"Document title\",\n    \"total_pages\": 100,\n    \"language\": \"zh/en\",\n    \"subject\": \"Subject area\"\n  },\n  \"chapters\": [\n    {\n      \"chapter_id\": \"1\",\n      \"title\": \"Chapter title\",\n      \"page_range\": [1, 20],\n      \"sections\": [\n        {\n          \"section_id\": \"1.1\",\n          \"title\": \"Section title\",\n          \"core_concepts\": [\"Concept 1\", \"Concept 2\"],\n          \"key_terms\": [\n            {\"term\": \"Term\", \"definition\": \"Definition\"}\n          ],\n          \"learning_objectives\": [\"Objective 1\", \"Objective 2\"],\n          \"difficulty\": \"easy/medium/hard\",\n          \"prerequisites\": [\"Prerequisite knowledge\"]\n        }\n      ]\n    }\n  ],\n  \"knowledge_graph\": {\n    \"nodes\": [\"Concept node list\"],\n    \"edges\": [{\"from\": \"Concept A\", \"to\": \"Concept B\", \"relation\": \"depends/contains/related\"}]\n  }\n}\n```\n\n**Parsing Rules:**\n1. Chapter Recognition — Identify hierarchy based on font size, bold, numbering, etc. Handle documents without clear chapter markers by logically segmenting.\n2. Concept Extraction — Identify bolded, highlighted, boxed important content. Extract proper nouns and term definitions. Identify formulas and theorems.\n3. Difficulty Assessment — Assess based on concept abstraction level, prerequisite knowledge, and content complexity.\n4. Quality Assurance — Ensure all chapters identified, verify knowledge point coverage completeness, check accuracy of concept definitions.\n\n### Phase 3: Generate Requested Formats\n\nBased on user-selected output formats, generate each in sequence. For each format, follow the corresponding section below.\n\n### Phase 4: Deliver Assets\n\nAfter all generation is complete:\n- Only return file paths, no previews allowed\n- No inline display of images/PDFs/audio/video in conversation\n- Audio/video files must not auto-play\n\nPresent to user using deliver_assets format:\n```\n<deliver_assets>\n<item>\n<path>file path</path>\n</item>\n</deliver_assets>\n```\n\n---\n\n## Supported Output Formats\n\n| Format | Output | Description |\n|--------|--------|-------------|\n| `notes` | `{topic}_notes.pdf` | Handwritten-style notes (annotated on original or generated from scratch) |\n| `quiz` | `{topic}_quiz.html` | Minimalist interactive HTML quiz with instant feedback |\n| `slides` | `{topic}_slides.pdf` + `{topic}_slides.pptx` | Visual slides |\n| `mindmap` | `{topic}_mindmap.png` + Mermaid text | Mind map image |\n| `audio` | `{topic}_audio.mp3` | Audio course in teacher-student dialogue format |\n| `all` | All of the above | Generate every format |\n\n---\n\n## Personalization: Grade Level Adaptation\n\nAll generated content must be adapted to the learner's grade level:\n\n| Grade | Language & Tone | Content Density | Visual Style |\n|-------|----------------|-----------------|--------------|\n| **Elementary** | Lively, simple Q&A, encouraging, story-style | Low density, more drawings, large font | Fun elements, bright colors, short text |\n| **Middle school** | Guided questioning, moderate challenges, youth-oriented | Moderate, image-text combination, clear labels | Image-text combination, moderate information |\n| **High school** | In-depth discussion, logical reasoning, appropriate academic tone | Higher density, logic diagrams | Professional feel, data visualization |\n| **University/Professional** | Seminar-style, critical thinking, professional terminology | High density, professional charts, complex structures | Academic style, comprehensive application |\n\n**Interest Adaptation** (applies to all formats):\n- Examples and metaphors use the user's interest field\n- Scenarios drawn from the user's familiar domain\n- Visual style and analogies match user interests\n\n---\n\n## Format 1: Notes Generation\n\n### Input Type Determination\n\n**Type A — Existing Paper/Courseware:**\n- PDF format academic papers, courseware/PPT exports, scanned textbook pages\n- Features: Fixed layout, page numbers, chapter numbering, formulas/charts\n- Action: Overlay handwritten notes on original pages\n\n**Type B — Non-existing Content:**\n- Plain text notes, knowledge point lists, oral transcripts, web content excerpts\n- Features: No fixed layout, needs reorganization\n- Action: Generate notes PDF from scratch\n\n### Type A Workflow: Adding Notes to Original Document\n\n**Step 1: Analyze Original Structure**\n\nAnalyze PDF content page by page:\n- Identify chapter titles and positions\n- Identify core concepts/terms\n- Identify formulas and their meanings\n- Identify problem/challenge statements\n- Identify solutions/methods\n- Identify key conclusions\n\n**Step 2: Plan Note Content**\n\nPlan handwritten annotations for each page (3-8 annotations per page, not too dense):\n\nAnnotation Types:\n1. **Chapter title translation/explanation** — e.g., original \"3.1 Preliminaries\" → annotate \"Background Knowledge\"\n2. **Key questions** — e.g., \"Key: How to reduce complexity?\"\n3. **Concept explanation** — e.g., annotate \"kernel trick\" next to formula\n4. **Problem marking** — e.g., \"Problem: memory overflow\"\n5. **Solutions** — e.g., \"Solution: forget gate\"\n6. **Formula notes** — e.g., \"recursive form\", \"write operation & read operation\"\n7. **Structure annotation** — e.g., use braces to mark formula groups, write \"→ O(N²) complexity\" beside\n\nAnnotation Planning Principles:\n- Positions avoid blocking key content\n- Utilize margins and paragraph gaps\n- Related content connected with lines or arrows\n\n**Step 3: Generate Annotated Images**\n\nConvert each PDF page to image, then use image generation tool to add handwritten-style annotations.\n\nHandwritten Annotation Style Requirements:\n- **Font:** Handwritten style, slightly tilted\n- **Color:** Unified colors throughout PDF, no more than 2\n  - Default: blue and pink (unless user specifies otherwise)\n  - All subsequent pages can only choose from these 2 colors\n  - Color assignment rules:\n    - Color 1 (blue/primary): Chapter titles, structure annotations, concept explanations, formula notes\n    - Color 2 (pink/accent): Key questions, problem marking, solutions\n- **Size:** Slightly larger than body text, eye-catching but not overwhelming\n- **Position:** Margins, paragraph gaps, blank space next to formulas\n\n**Step 4: Compile PDF**\n- Maintain original page order\n- Image quality: 150 DPI\n- Compression quality: 90%\n\n### Type B Workflow: Generating Notes from Scratch\n\n**Step 1: Organize Content Structure**\n- Main title → Chapters/modules → Core concepts → Key points/details → Examples/applications\n\n**Step 2: Design Note Layout**\n\nLayout Elements:\n- Title area: Large handwritten title\n- Body area: Handwritten-style bullet points\n- Diagram area: Concept maps, flowcharts, relationship diagrams (hand-drawn style)\n- Annotation area: Key markers, question marks, exclamation marks\n- Blank area: Space reserved for user's own notes\n\n**Step 3: Generate Note Page Images**\n\nEach page contains:\n- Page title (handwritten large text)\n- Core content (handwritten bullet points)\n- Diagrams (hand-drawn style concept maps/flowcharts)\n- Key annotations (boxes, arrows, underlines)\n- Notes (like \"Important!\", \"Common mistake\", \"Remember this\")\n\nStyle Requirements:\n- **Overall:** Looks like carefully made student notes, not printed document\n- **Font:** Handwritten, varying sizes (large for titles, medium for body, small for notes)\n- **Color:** Unified colors throughout PDF, no more than 2\n  - Default: blue and pink (unless user specifies otherwise)\n  - Color assignment: Blue (titles, framework, notes), Pink (key points)\n- **Layout:** Organized but not rigid, slight tilting and variation allowed\n- **Elements:** Arrows, underlines, boxes, cloud frames, asterisks — use only when necessary\n\n**Step 4: Compile PDF**\n- Arrange in logical content order\n- Image quality: 150 DPI, compression quality: 90%\n\n### Notes Output\n- File: `{topic}_notes.pdf`\n- Only return file path, no preview in conversation\n- Do not output intermediate image files or content scripts\n\n### Notes Quality Standards\n1. **Content Accuracy** — Annotations based on original text; translation/explanation accurate; no added information\n2. **Annotation Value** — Annotations help understanding, not simple repetition; key points highlight important concepts; problems and solutions correspond clearly\n3. **Visual Effect** — Handwritten style natural, not machine-printed; color coordination harmonious; annotation positions reasonable\n4. **Usability** — PDF printable; suitable for screen reading; reasonable file size\n\n---\n\n## Format 2: Quiz Generation\n\n### Question Design\n\nAt least 5 questions per section. Distribution:\n- Multiple choice (multiple_choice): 2-3 questions\n- True/false (true_false): 1-2 questions\n- Fill in the blank (fill_blank): 1-2 questions\n\nDifficulty distribution:\n- 40% Easy (memory, comprehension)\n- 40% Medium (application)\n- 20% Hard (analysis, synthesis)\n\nEach question must include:\n- Question content (using personalized scenario)\n- Correct answer\n- Answer explanation (has teaching value, not just \"the answer is X\")\n- Related core concept\n\n### HTML Generation\n\nGenerate a single HTML file containing all questions and interaction logic.\n\n**Design Principle: Minimalist**\n\nVisual Style:\n- Pure white background\n- Black text\n- No decorative elements, no icons, no gradients, no shadows\n- No borders or only 1px gray thin lines\n- Font: System default font\n- Minimal CSS, no UI frameworks\n\nInteraction Design:\n- Click option to select, selected state distinguished by slight background color\n- Show correct/incorrect and explanation immediately after submit\n- Correct: Green text \"Correct\"\n- Incorrect: Red text \"Incorrect\" + correct answer + explanation\n- Show total score at end\n\n**HTML Structure Template:**\n\n```html\n<!DOCTYPE html>\n<html lang=\"en\">\n<head>\n  <meta charset=\"UTF-8\">\n  <meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">\n  <title>Chapter Quiz</title>\n  <style>\n    body {\n      font-family: system-ui, sans-serif;\n      max-width: 600px;\n      margin: 40px auto;\n      padding: 20px;\n      line-height: 1.6;\n    }\n    h1 { font-size: 1.5em; font-weight: normal; }\n    .question { margin: 30px 0; }\n    .question-text { margin-bottom: 15px; }\n    .option {\n      display: block;\n      padding: 10px;\n      margin: 5px 0;\n      cursor: pointer;\n    }\n    .option:hover { background: #f5f5f5; }\n    .option.selected { background: #e8e8e8; }\n    .feedback { margin-top: 10px; font-size: 0.9em; }\n    .correct { color: #2e7d32; }\n    .incorrect { color: #c62828; }\n    .explanation { color: #666; margin-top: 5px; }\n    button {\n      padding: 10px 20px;\n      background: #333;\n      color: white;\n      border: none;\n      cursor: pointer;\n      margin-top: 20px;\n    }\n    .score { font-size: 1.2em; margin-top: 30px; }\n  </style>\n</head>\n<body>\n  <h1>Chapter Title - Quiz</h1>\n\n  <div class=\"question\" data-answer=\"A\">\n    <div class=\"question-text\">1. Question content</div>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"A\"> A. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"B\"> B. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"C\"> C. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"D\"> D. Option</label>\n    <div class=\"feedback\"></div>\n  </div>\n\n  <!-- More questions... -->\n\n  <button onclick=\"submit()\">Submit</button>\n  <div class=\"score\"></div>\n\n  <script>\n    const explanations = {\n      q1: \"Explanation content...\",\n      // ...\n    };\n\n    function submit() {\n      let correct = 0;\n      document.querySelectorAll('.question').forEach((q, i) => {\n        const answer = q.dataset.answer;\n        const selected = q.querySelector('input:checked');\n        const feedback = q.querySelector('.feedback');\n        const qName = 'q' + (i + 1);\n\n        if (selected && selected.value === answer) {\n          feedback.innerHTML = '<span class=\"correct\">Correct</span>';\n          correct++;\n        } else {\n          feedback.innerHTML = '<span class=\"incorrect\">Incorrect</span> Correct answer: ' + answer +\n            '<div class=\"explanation\">' + explanations[qName] + '</div>';\n        }\n      });\n\n      document.querySelector('.score').textContent =\n        'Score: ' + correct + '/' + document.querySelectorAll('.question').length;\n    }\n  </script>\n</body>\n</html>\n```\n\n### Quiz Output\n- File: `{topic}_quiz.html`\n- Only return file path, no preview in conversation\n- Do not output JSON data, CSS files, or JS files separately\n\n### Quiz Quality Standards\n1. **Content Accuracy** — All knowledge points based on original textbook; answers and explanations correct; question wording clear and unambiguous\n2. **Personalization** — Question scenarios match user interests; difficulty matches grade level; language style suits target audience\n3. **Interaction Experience** — Click response instant; feedback clear; explanations have teaching value\n4. **Visual Minimalism** — No decorative elements; no framework dependencies; file size minimized\n\n---\n\n## Format 3: Slides Generation\n\n### Design Considerations\n\nTreat these as a flexible menu, not a mandatory checklist:\n\n1. **Topic, Purpose & Audience** — What is this about? Who needs to understand it? Where will it be presented?\n2. **Content Foundation & Sources** — What materials or data need to be presented?\n3. **Visual Approach (CRITICAL)**\n   - Default to explanatory visuals: cutaway views, annotated structure diagrams, exploded views, schematic illustrations\n   - Visual elements are primary information carriers, not decorative backgrounds for text lists\n   - Default information density matches professional infographics and technical illustrations\n   - **CRITICAL:** Diagrams must convey information through structure, not just provide atmosphere. Text should be labels/annotations, not main content. Reject purely decorative visuals with core information dependent on text lists\n   - Reject the inefficient pattern of \"large white space + centered single line of text\"\n4. **Narrative Flow & Chapters** — How should viewers move through the content? How is slide flow arranged?\n5. **Text Style & Density**\n   - Language: Explanatory text uses language explicitly requested by user, otherwise match user's conversation language\n   - Typography: Chinese and English titles preferably use serif fonts (Chinese uses Song font family)\n6. **Visual Style, Color & Mood**\n   - Visual language of encyclopedias and reference books: explanatory diagrams, cutaway illustrations, annotated structures\n   - Refined spatial composition and typographic precision of high-end journals\n   - Intentional asymmetry and layered information design of contemporary design publications\n   - Apply asymmetric grids, intentional breathing space, layered information organization, diagonal composition, dynamic typography as internalized design language\n   - **Color restriction:** Unless user explicitly specifies, do NOT use blue or purple as theme color or background color\n\n### Slides Workflow\n\n**Step 1: Design Strategy — Create Content Script**\n\nInformation architecture first: Structure content into hierarchical slides, each slide as an information unit defined by what data/facts/relationships it carries. Let content volume naturally determine slide count.\n\nOutput `content_script.md`:\n\n```\n# Slides Content Script\n\n## Slide 1: [Title]\n**Subtopic A**: [Label]\n[50-80 word narrative paragraph describing information content to be visualized]\n\n**Subtopic B**: [Label]\n[50-80 word narrative paragraph]\n\n## Slide 2: [Title]\n...\n```\n\nContent Script Specification:\n- Only describe \"what information needs to be presented\", not \"how to present it\"\n- Do NOT include \"Visual Description\" sections\n- Do NOT describe colors, backgrounds, decorative elements, atmosphere effects, mood, or layout details\n- Focus on pure information architecture\n- 2-3 focused subtopics per slide\n\n**Step 2: Sequential Image Generation**\n\nUse image generation tool to generate slides one by one:\n- First slide: Use gen_images (create from scratch)\n- Subsequent slides: Use edit_images, base_image_file points to previous slide\n\nFormat: Default 16:9 landscape ratio. Save each slide image locally.\n\n**Prompt Construction for Each Slide — Must include these 6 points:**\n\n1. **Visualization Type** — Prioritize diagram forms over text-dominated presentations: cutaway views, flowcharts, annotated structure diagrams, relationship diagrams, timeline overlays. Integrate multiple subtopics into unified visual structure. Avoid \"parallel cards/grid displays/multi-column layouts\" and text-heavy traditional typography.\n\n2. **Information Hierarchy** — Primary and secondary information distinguished through visual hierarchy (size, position, contrast). Not flat lists.\n\n3. **Composition Instructions** — Asymmetric layout, diagonal momentum, and other methods to break rigid symmetry.\n\n4. **Density Requirements** — Clear information hierarchy over quantity. Appropriate white space serves readability, but not empty and sparse.\n\n5. **Layout Independence** — Explicitly state this slide's visualization type is chosen based on its content, not copying previous slide. Re-evaluate what this specific content needs. But describe inherited elements in detail.\n\n6. **Style Consistency** — If user provided visual style or reference images, each prompt must describe that style's characteristics in detail.\n\n**Step 3: Compile Output**\n\nAfter generating all slide images:\n- Auto-compile into PDF (150 DPI, 95% quality, controlled file size)\n- Auto-compile into PPTX presentation\n\n### Slides Output\n- Files: `{topic}_slides.pdf` + `{topic}_slides.pptx`\n- Only return file paths, no preview in conversation\n- Do not output individual slide images, summary documents, content outlines, design descriptions, or usage instructions\n\n---\n\n## Format 4: Mind Map Generation\n\n### Mind Map Workflow\n\n**Step 1: Design Content Structure**\n\nDetermine node hierarchy and relationships:\n- Root node: Chapter theme\n- Level 1 nodes: Core concepts\n- Level 2 nodes: Detail points\n- No more than 4 levels\n- Each node text concise (no more than 10 characters)\n- Mark relationships between concepts (parallel/progressive/causal/contrast)\n\n**Step 2: Generate Image**\n\nUse gen_images to generate mind map image:\n- Format: 16:9 or square (based on content)\n- Style: Clear visual hierarchy, professional infographic style\n\n**Step 3: Output**\n\n- Mind map image: `{topic}_mindmap.png`\n- Attached Mermaid format text (optional, for users who need to edit)\n- Only return file path, no image preview in conversation\n\n---\n\n## Format 5: Audio Course Generation\n\n### Audio Workflow\n\n**Step 1: Write Dialogue Script**\n\nWrite teacher-student dialogue script:\n\n```\nOpening (about 1 minute)\n- Teacher greets, introduces today's topic\n- Student responds, expresses existing knowledge or questions\n- Teacher builds connection using user's interest field\n\nPart One: Concept Introduction (about 4 minutes)\n- Teacher asks questions from user's interest scenario\n- Student observes/answers\n- Teacher introduces core concept, defines in conversational manner\n- Student requests examples\n- Teacher explains in detail with personalized examples\n- Student restates in own words to confirm understanding\n\nPart Two: Deep Understanding (about 5 minutes)\n- Teacher explains important characteristics of concept\n- Student raises common confusion/misconception\n- Teacher clarifies misconception\n- Student poses hypothetical questions\n- Teacher answers and extends\n\nPart Three: Application Practice (about 3 minutes)\n- Teacher gives question\n- Student thinks and answers\n- Teacher provides feedback (affirmation or guidance)\n\nSummary (about 2 minutes)\n- Student attempts to summarize what was learned\n- Teacher supplements and affirms\n- Student expresses gains, connects to practical application\n- Exchange farewells\n```\n\nScript Requirements:\n- Dialogue natural, matches real teacher-student conversation rhythm\n- Avoid written expression\n- Include interjections (\"um\", \"well\", \"oh right\")\n- Allow student to \"interrupt\" with questions\n- All examples sourced from user's interest field\n- About 150-180 words per minute\n\n### Character Settings\n\n**Teacher Character:**\n- Professional yet approachable\n- Good at using metaphors to explain complex concepts\n- Patient in answering questions\n- Timely encouragement and affirmation\n\n**Student Character:**\n- Curious, actively asks questions\n- Represents target user's perspective\n- Makes common mistakes, raises typical confusions\n- Has own interest background (consistent with user settings)\n\n**Step 2: Generate Audio**\n\nUse audio generation tool to convert script to audio:\n- Teacher voice: Warm, professional, patient\n- Student voice: Curious, lively, sincere\n- Speed: Medium for concept explanation, natural rhythm for dialogue, slightly faster for summary\n\n**Step 3: Output**\n- File: `{topic}_audio.mp3`\n- Only return file path, no preview or playback in conversation\n- No auto-play\n- Do not output script files or production notes\n\n### Audio Quality Standards\n1. **Listening Experience** — Sounds like real conversation, not script reading; rhythm varies; key content emphasized\n2. **Learning Effect** — Concept explanation clear; student questions represent real confusion; practice section has testing effect\n3. **Personalization** — Examples 100% from user's interest field; student character gives user identification; language style matches grade\n4. **Audio Quality** — Clear sound; duration about 15 minutes; directly playable\n\n---\n\n## Critical Constraints\n\n1. **Content Fidelity** — All content must be based on original textbook/source material. No unverified information added.\n2. **Grade Adaptation** — Adjust content depth and expression based on grade level for ALL formats.\n3. **Output Rules** — Only return file paths. No inline display of images/PDFs/audio/video. No auto-play. No intermediate files.\n4. **Color Constraints (Notes)** — Maximum 2 colors per PDF. Default blue + pink.\n5. **Color Constraints (Slides)** — Do NOT use blue or purple as theme/background color unless user explicitly requests.\n6. **Image Quality** — Notes: 150 DPI, 90% compression. Slides: 150 DPI, 95% quality.\n7. **Mind Map Depth** — No more than 4 levels. Node text no more than 10 characters.\n8. **Quiz Minimalism** — No UI frameworks, no decorative elements, system default font only.\n\n## Common Mistakes to Avoid\n\n1. **Adding unverified information** — Stick to the source material only\n2. **Ignoring grade level** — Elementary content should not use university-level terminology\n3. **Previewing outputs in conversation** — Never display images, PDFs, or play audio inline\n4. **Dense annotations on notes** — Keep 3-8 annotations per page, not more\n5. **Decorative slides** — Visuals must convey information through structure, not just atmosphere\n6. **Text-heavy slides** — Diagrams should be primary carriers, not text lists with decorative backgrounds\n7. **Using blue/purple in slides** — Forbidden unless user explicitly requests\n8. **Flat quiz feedback** — \"The answer is X\" has no teaching value; always explain why\n9. **Robotic audio dialogue** — Must sound like natural conversation with interjections and interruptions\n10. **Outputting intermediate files** — Only deliver final output file paths\n\n## File & Output Conventions\n\n| Format | Filename Pattern | File Type |\n|--------|-----------------|-----------|\n| Notes | `{topic}_notes.pdf` | PDF |\n| Quiz | `{topic}_quiz.html` | HTML |\n| Slides | `{topic}_slides.pdf`, `{topic}_slides.pptx` | PDF, PPTX |\n| Mind Map | `{topic}_mindmap.png` | PNG |\n| Audio | `{topic}_audio.mp3` | MP3 |\n\nAll files use the topic name as prefix. Deliver all outputs together using `<deliver_assets>` format after all generation is complete.\n\nFile v1.2.0:_meta.json\n\n{\n  \"ownerId\": \"kn712egma3976f4b68hksznyrs85wsr5\",\n  \"slug\": \"eric-knowledge-digest\",\n  \"version\": \"1.2.0\",\n  \"publishedAt\": 1777976261510\n}\n\nArchive v1.0.5: 2 files, 10471 bytes\n\nFiles: _meta.json (140b), SKILL.md (26234b)\n\nFile v1.0.5:SKILL.md\n\n---\nname: knowledge-digest\ndescription: \"Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, and mind maps. Trigger: learning materials, convert textbook, study notes, quiz generation, slides from PDF, mind map, audio course.\"\n---\n\n# KnowledgeDigest — Unified Learning Content Converter\n\n## Overview\n\nKnowledgeDigest converts textbooks, PDFs, or topic descriptions into personalized, multimodal learning experiences. It analyzes source content, then generates any combination of: handwritten-style notes (PDF), interactive quiz webpages (HTML), slides (PDF+PPTX), mind maps (image+Mermaid), and audio courses (MP3). All output is adapted to the learner's grade level and interests.\n\n## Workflow\n\n### Phase 1: Gather User Input\n\n1. Identify what the user has provided:\n   - Uploaded PDF/textbook file (optional)\n   - Topic/direction description\n   - Grade level (elementary / middle school / high school / university / professional)\n   - Expected output format(s)\n\n2. If no PDF/textbook uploaded and no source materials specified (only topic/direction provided):\n   - Ask user:\n     - **Option A:** \"I have materials, uploading now\"\n     - **Option B:** \"No materials, please search and generate courseware about [topic]\"\n   - If user selects B:\n     - Use search tools to collect authoritative materials on the topic\n     - Organize into structured content, generate a basic courseware PDF\n     - Send PDF to user for confirmation: \"This is the basic material I compiled for [topic], please confirm if usable?\"\n     - Continue after user confirmation\n\n3. **Default output formats** (if user does not specify): mindmap + slides (PDF only) + quiz\n\n### Phase 2: Content Analysis\n\nParse the PDF or structured content to extract:\n\n**Document Parsing:**\n- Identify chapter structure (chapters, sections, subsections)\n- Extract heading hierarchy and table of contents\n- Identify body text, images, tables, formulas, and other elements\n\n**Core Concept Extraction:**\n- Identify core concepts and key terms in each chapter\n- Extract definitions, theorems, formulas, and important content\n- Mark difficult points and key knowledge\n\n**Learning Objective Analysis:**\n- Infer learning objectives for each chapter\n- Identify prerequisite knowledge requirements\n- Analyze dependencies between knowledge points\n\n**Output structured analysis results in this format:**\n\n```json\n{\n  \"document_info\": {\n    \"title\": \"Document title\",\n    \"total_pages\": 100,\n    \"language\": \"zh/en\",\n    \"subject\": \"Subject area\"\n  },\n  \"chapters\": [\n    {\n      \"chapter_id\": \"1\",\n      \"title\": \"Chapter title\",\n      \"page_range\": [1, 20],\n      \"sections\": [\n        {\n          \"section_id\": \"1.1\",\n          \"title\": \"Section title\",\n          \"core_concepts\": [\"Concept 1\", \"Concept 2\"],\n          \"key_terms\": [\n            {\"term\": \"Term\", \"definition\": \"Definition\"}\n          ],\n          \"learning_objectives\": [\"Objective 1\", \"Objective 2\"],\n          \"difficulty\": \"easy/medium/hard\",\n          \"prerequisites\": [\"Prerequisite knowledge\"]\n        }\n      ]\n    }\n  ],\n  \"knowledge_graph\": {\n    \"nodes\": [\"Concept node list\"],\n    \"edges\": [{\"from\": \"Concept A\", \"to\": \"Concept B\", \"relation\": \"depends/contains/related\"}]\n  }\n}\n```\n\n**Parsing Rules:**\n1. Chapter Recognition — Identify hierarchy based on font size, bold, numbering, etc. Handle documents without clear chapter markers by logically segmenting.\n2. Concept Extraction — Identify bolded, highlighted, boxed important content. Extract proper nouns and term definitions. Identify formulas and theorems.\n3. Difficulty Assessment — Assess based on concept abstraction level, prerequisite knowledge, and content complexity.\n4. Quality Assurance — Ensure all chapters identified, verify knowledge point coverage completeness, check accuracy of concept definitions.\n\n### Phase 3: Generate Requested Formats\n\nBased on user-selected output formats, generate each in sequence. For each format, follow the corresponding section below.\n\n### Phase 4: Deliver Assets\n\nAfter all generation is complete:\n- Only return file paths, no previews allowed\n- No inline display of images/PDFs/audio/video in conversation\n- Audio/video files must not auto-play\n\nPresent to user using deliver_assets format:\n```\n<deliver_assets>\n<item>\n<path>file path</path>\n</item>\n</deliver_assets>\n```\n\n---\n\n## Supported Output Formats\n\n| Format | Output | Description |\n|--------|--------|-------------|\n| `notes` | `{topic}_notes.pdf` | Handwritten-style notes (annotated on original or generated from scratch) |\n| `quiz` | `{topic}_quiz.html` | Minimalist interactive HTML quiz with instant feedback |\n| `slides` | `{topic}_slides.pdf` + `{topic}_slides.pptx` | Visual slides |\n| `mindmap` | `{topic}_mindmap.png` + Mermaid text | Mind map image |\n| `audio` | `{topic}_audio.mp3` | Audio course in teacher-student dialogue format |\n| `all` | All of the above | Generate every format |\n\n---\n\n## Personalization: Grade Level Adaptation\n\nAll generated content must be adapted to the learner's grade level:\n\n| Grade | Language & Tone | Content Density | Visual Style |\n|-------|----------------|-----------------|--------------|\n| **Elementary** | Lively, simple Q&A, encouraging, story-style | Low density, more drawings, large font | Fun elements, bright colors, short text |\n| **Middle school** | Guided questioning, moderate challenges, youth-oriented | Moderate, image-text combination, clear labels | Image-text combination, moderate information |\n| **High school** | In-depth discussion, logical reasoning, appropriate academic tone | Higher density, logic diagrams | Professional feel, data visualization |\n| **University/Professional** | Seminar-style, critical thinking, professional terminology | High density, professional charts, complex structures | Academic style, comprehensive application |\n\n**Interest Adaptation** (applies to all formats):\n- Examples and metaphors use the user's interest field\n- Scenarios drawn from the user's familiar domain\n- Visual style and analogies match user interests\n\n---\n\n## Format 1: Notes Generation\n\n### Input Type Determination\n\n**Type A — Existing Paper/Courseware:**\n- PDF format academic papers, courseware/PPT exports, scanned textbook pages\n- Features: Fixed layout, page numbers, chapter numbering, formulas/charts\n- Action: Overlay handwritten notes on original pages\n\n**Type B — Non-existing Content:**\n- Plain text notes, knowledge point lists, oral transcripts, web content excerpts\n- Features: No fixed layout, needs reorganization\n- Action: Generate notes PDF from scratch\n\n### Type A Workflow: Adding Notes to Original Document\n\n**Step 1: Analyze Original Structure**\n\nAnalyze PDF content page by page:\n- Identify chapter titles and positions\n- Identify core concepts/terms\n- Identify formulas and their meanings\n- Identify problem/challenge statements\n- Identify solutions/methods\n- Identify key conclusions\n\n**Step 2: Plan Note Content**\n\nPlan handwritten annotations for each page (3-8 annotations per page, not too dense):\n\nAnnotation Types:\n1. **Chapter title translation/explanation** — e.g., original \"3.1 Preliminaries\" → annotate \"Background Knowledge\"\n2. **Key questions** — e.g., \"Key: How to reduce complexity?\"\n3. **Concept explanation** — e.g., annotate \"kernel trick\" next to formula\n4. **Problem marking** — e.g., \"Problem: memory overflow\"\n5. **Solutions** — e.g., \"Solution: forget gate\"\n6. **Formula notes** — e.g., \"recursive form\", \"write operation & read operation\"\n7. **Structure annotation** — e.g., use braces to mark formula groups, write \"→ O(N²) complexity\" beside\n\nAnnotation Planning Principles:\n- Positions avoid blocking key content\n- Utilize margins and paragraph gaps\n- Related content connected with lines or arrows\n\n**Step 3: Generate Annotated Images**\n\nConvert each PDF page to image, then use image generation tool to add handwritten-style annotations.\n\nHandwritten Annotation Style Requirements:\n- **Font:** Handwritten style, slightly tilted\n- **Color:** Unified colors throughout PDF, no more than 2\n  - Default: blue and pink (unless user specifies otherwise)\n  - All subsequent pages can only choose from these 2 colors\n  - Color assignment rules:\n    - Color 1 (blue/primary): Chapter titles, structure annotations, concept explanations, formula notes\n    - Color 2 (pink/accent): Key questions, problem marking, solutions\n- **Size:** Slightly larger than body text, eye-catching but not overwhelming\n- **Position:** Margins, paragraph gaps, blank space next to formulas\n\n**Step 4: Compile PDF**\n- Maintain original page order\n- Image quality: 150 DPI\n- Compression quality: 90%\n\n### Type B Workflow: Generating Notes from Scratch\n\n**Step 1: Organize Content Structure**\n- Main title → Chapters/modules → Core concepts → Key points/details → Examples/applications\n\n**Step 2: Design Note Layout**\n\nLayout Elements:\n- Title area: Large handwritten title\n- Body area: Handwritten-style bullet points\n- Diagram area: Concept maps, flowcharts, relationship diagrams (hand-drawn style)\n- Annotation area: Key markers, question marks, exclamation marks\n- Blank area: Space reserved for user's own notes\n\n**Step 3: Generate Note Page Images**\n\nEach page contains:\n- Page title (handwritten large text)\n- Core content (handwritten bullet points)\n- Diagrams (hand-drawn style concept maps/flowcharts)\n- Key annotations (boxes, arrows, underlines)\n- Notes (like \"Important!\", \"Common mistake\", \"Remember this\")\n\nStyle Requirements:\n- **Overall:** Looks like carefully made student notes, not printed document\n- **Font:** Handwritten, varying sizes (large for titles, medium for body, small for notes)\n- **Color:** Unified colors throughout PDF, no more than 2\n  - Default: blue and pink (unless user specifies otherwise)\n  - Color assignment: Blue (titles, framework, notes), Pink (key points)\n- **Layout:** Organized but not rigid, slight tilting and variation allowed\n- **Elements:** Arrows, underlines, boxes, cloud frames, asterisks — use only when necessary\n\n**Step 4: Compile PDF**\n- Arrange in logical content order\n- Image quality: 150 DPI, compression quality: 90%\n\n### Notes Output\n- File: `{topic}_notes.pdf`\n- Only return file path, no preview in conversation\n- Do not output intermediate image files or content scripts\n\n### Notes Quality Standards\n1. **Content Accuracy** — Annotations based on original text; translation/explanation accurate; no added information\n2. **Annotation Value** — Annotations help understanding, not simple repetition; key points highlight important concepts; problems and solutions correspond clearly\n3. **Visual Effect** — Handwritten style natural, not machine-printed; color coordination harmonious; annotation positions reasonable\n4. **Usability** — PDF printable; suitable for screen reading; reasonable file size\n\n---\n\n## Format 2: Quiz Generation\n\n### Question Design\n\nAt least 5 questions per section. Distribution:\n- Multiple choice (multiple_choice): 2-3 questions\n- True/false (true_false): 1-2 questions\n- Fill in the blank (fill_blank): 1-2 questions\n\nDifficulty distribution:\n- 40% Easy (memory, comprehension)\n- 40% Medium (application)\n- 20% Hard (analysis, synthesis)\n\nEach question must include:\n- Question content (using personalized scenario)\n- Correct answer\n- Answer explanation (has teaching value, not just \"the answer is X\")\n- Related core concept\n\n### HTML Generation\n\nGenerate a single HTML file containing all questions and interaction logic.\n\n**Design Principle: Minimalist**\n\nVisual Style:\n- Pure white background\n- Black text\n- No decorative elements, no icons, no gradients, no shadows\n- No borders or only 1px gray thin lines\n- Font: System default font\n- Minimal CSS, no UI frameworks\n\nInteraction Design:\n- Click option to select, selected state distinguished by slight background color\n- Show correct/incorrect and explanation immediately after submit\n- Correct: Green text \"Correct\"\n- Incorrect: Red text \"Incorrect\" + correct answer + explanation\n- Show total score at end\n\n**HTML Structure Template:**\n\n```html\n<!DOCTYPE html>\n<html lang=\"en\">\n<head>\n  <meta charset=\"UTF-8\">\n  <meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">\n  <title>Chapter Quiz</title>\n  <style>\n    body {\n      font-family: system-ui, sans-serif;\n      max-width: 600px;\n      margin: 40px auto;\n      padding: 20px;\n      line-height: 1.6;\n    }\n    h1 { font-size: 1.5em; font-weight: normal; }\n    .question { margin: 30px 0; }\n    .question-text { margin-bottom: 15px; }\n    .option {\n      display: block;\n      padding: 10px;\n      margin: 5px 0;\n      cursor: pointer;\n    }\n    .option:hover { background: #f5f5f5; }\n    .option.selected { background: #e8e8e8; }\n    .feedback { margin-top: 10px; font-size: 0.9em; }\n    .correct { color: #2e7d32; }\n    .incorrect { color: #c62828; }\n    .explanation { color: #666; margin-top: 5px; }\n    button {\n      padding: 10px 20px;\n      background: #333;\n      color: white;\n      border: none;\n      cursor: pointer;\n      margin-top: 20px;\n    }\n    .score { font-size: 1.2em; margin-top: 30px; }\n  </style>\n</head>\n<body>\n  <h1>Chapter Title - Quiz</h1>\n\n  <div class=\"question\" data-answer=\"A\">\n    <div class=\"question-text\">1. Question content</div>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"A\"> A. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"B\"> B. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"C\"> C. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"D\"> D. Option</label>\n    <div class=\"feedback\"></div>\n  </div>\n\n  <!-- More questions... -->\n\n  <button onclick=\"submit()\">Submit</button>\n  <div class=\"score\"></div>\n\n  <script>\n    const explanations = {\n      q1: \"Explanation content...\",\n      // ...\n    };\n\n    function submit() {\n      let correct = 0;\n      document.querySelectorAll('.question').forEach((q, i) => {\n        const answer = q.dataset.answer;\n        const selected = q.querySelector('input:checked');\n        const feedback = q.querySelector('.feedback');\n        const qName = 'q' + (i + 1);\n\n        if (selected && selected.value === answer) {\n          feedback.innerHTML = '<span class=\"correct\">Correct</span>';\n          correct++;\n        } else {\n          feedback.innerHTML = '<span class=\"incorrect\">Incorrect</span> Correct answer: ' + answer +\n            '<div class=\"explanation\">' + explanations[qName] + '</div>';\n        }\n      });\n\n      document.querySelector('.score').textContent =\n        'Score: ' + correct + '/' + document.querySelectorAll('.question').length;\n    }\n  </script>\n</body>\n</html>\n```\n\n### Quiz Output\n- File: `{topic}_quiz.html`\n- Only return file path, no preview in conversation\n- Do not output JSON data, CSS files, or JS files separately\n\n### Quiz Quality Standards\n1. **Content Accuracy** — All knowledge points based on original textbook; answers and explanations correct; question wording clear and unambiguous\n2. **Personalization** — Question scenarios match user interests; difficulty matches grade level; language style suits target audience\n3. **Interaction Experience** — Click response instant; feedback clear; explanations have teaching value\n4. **Visual Minimalism** — No decorative elements; no framework dependencies; file size minimized\n\n---\n\n## Format 3: Slides Generation\n\n### Design Considerations\n\nTreat these as a flexible menu, not a mandatory checklist:\n\n1. **Topic, Purpose & Audience** — What is this about? Who needs to understand it? Where will it be presented?\n2. **Content Foundation & Sources** — What materials or data need to be presented?\n3. **Visual Approach (CRITICAL)**\n   - Default to explanatory visuals: cutaway views, annotated structure diagrams, exploded views, schematic illustrations\n   - Visual elements are primary information carriers, not decorative backgrounds for text lists\n   - Default information density matches professional infographics and technical illustrations\n   - **CRITICAL:** Diagrams must convey information through structure, not just provide atmosphere. Text should be labels/annotations, not main content. Reject purely decorative visuals with core information dependent on text lists\n   - Reject the inefficient pattern of \"large white space + centered single line of text\"\n4. **Narrative Flow & Chapters** — How should viewers move through the content? How is slide flow arranged?\n5. **Text Style & Density**\n   - Language: Explanatory text uses language explicitly requested by user, otherwise match user's conversation language\n   - Typography: Chinese and English titles preferably use serif fonts (Chinese uses Song font family)\n6. **Visual Style, Color & Mood**\n   - Visual language of encyclopedias and reference books: explanatory diagrams, cutaway illustrations, annotated structures\n   - Refined spatial composition and typographic precision of high-end journals\n   - Intentional asymmetry and layered information design of contemporary design publications\n   - Apply asymmetric grids, intentional breathing space, layered information organization, diagonal composition, dynamic typography as internalized design language\n   - **Color restriction:** Unless user explicitly specifies, do NOT use blue or purple as theme color or background color\n\n### Slides Workflow\n\n**Step 1: Design Strategy — Create Content Script**\n\nInformation architecture first: Structure content into hierarchical slides, each slide as an information unit defined by what data/facts/relationships it carries. Let content volume naturally determine slide count.\n\nOutput `content_script.md`:\n\n```\n# Slides Content Script\n\n## Slide 1: [Title]\n**Subtopic A**: [Label]\n[50-80 word narrative paragraph describing information content to be visualized]\n\n**Subtopic B**: [Label]\n[50-80 word narrative paragraph]\n\n## Slide 2: [Title]\n...\n```\n\nContent Script Specification:\n- Only describe \"what information needs to be presented\", not \"how to present it\"\n- Do NOT include \"Visual Description\" sections\n- Do NOT describe colors, backgrounds, decorative elements, atmosphere effects, mood, or layout details\n- Focus on pure information architecture\n- 2-3 focused subtopics per slide\n\n**Step 2: Sequential Image Generation**\n\nUse image generation tool to generate slides one by one:\n- First slide: Use gen_images (create from scratch)\n- Subsequent slides: Use edit_images, base_image_file points to previous slide\n\nFormat: Default 16:9 landscape ratio. Save each slide image locally.\n\n**Prompt Construction for Each Slide — Must include these 6 points:**\n\n1. **Visualization Type** — Prioritize diagram forms over text-dominated presentations: cutaway views, flowcharts, annotated structure diagrams, relationship diagrams, timeline overlays. Integrate multiple subtopics into unified visual structure. Avoid \"parallel cards/grid displays/multi-column layouts\" and text-heavy traditional typography.\n\n2. **Information Hierarchy** — Primary and secondary information distinguished through visual hierarchy (size, position, contrast). Not flat lists.\n\n3. **Composition Instructions** — Asymmetric layout, diagonal momentum, and other methods to break rigid symmetry.\n\n4. **Density Requirements** — Clear information hierarchy over quantity. Appropriate white space serves readability, but not empty and sparse.\n\n5. **Layout Independence** — Explicitly state this slide's visualization type is chosen based on its content, not copying previous slide. Re-evaluate what this specific content needs. But describe inherited elements in detail.\n\n6. **Style Consistency** — If user provided visual style or reference images, each prompt must describe that style's characteristics in detail.\n\n**Step 3: Compile Output**\n\nAfter generating all slide images:\n- Auto-compile into PDF (150 DPI, 95% quality, controlled file size)\n- Auto-compile into PPTX presentation\n\n### Slides Output\n- Files: `{topic}_slides.pdf` + `{topic}_slides.pptx`\n- Only return file paths, no preview in conversation\n- Do not output individual slide images, summary documents, content outlines, design descriptions, or usage instructions\n\n---\n\n## Format 4: Mind Map Generation\n\n### Mind Map Workflow\n\n**Step 1: Design Content Structure**\n\nDetermine node hierarchy and relationships:\n- Root node: Chapter theme\n- Level 1 nodes: Core concepts\n- Level 2 nodes: Detail points\n- No more than 4 levels\n- Each node text concise (no more than 10 characters)\n- Mark relationships between concepts (parallel/progressive/causal/contrast)\n\n**Step 2: Generate Image**\n\nUse gen_images to generate mind map image:\n- Format: 16:9 or square (based on content)\n- Style: Clear visual hierarchy, professional infographic style\n\n**Step 3: Output**\n\n- Mind map image: `{topic}_mindmap.png`\n- Attached Mermaid format text (optional, for users who need to edit)\n- Only return file path, no image preview in conversation\n\n---\n\n## Format 5: Audio Course Generation\n\n### Audio Workflow\n\n**Step 1: Write Dialogue Script**\n\nWrite teacher-student dialogue script:\n\n```\nOpening (about 1 minute)\n- Teacher greets, introduces today's topic\n- Student responds, expresses existing knowledge or questions\n- Teacher builds connection using user's interest field\n\nPart One: Concept Introduction (about 4 minutes)\n- Teacher asks questions from user's interest scenario\n- Student observes/answers\n- Teacher introduces core concept, defines in conversational manner\n- Student requests examples\n- Teacher explains in detail with personalized examples\n- Student restates in own words to confirm understanding\n\nPart Two: Deep Understanding (about 5 minutes)\n- Teacher explains important characteristics of concept\n- Student raises common confusion/misconception\n- Teacher clarifies misconception\n- Student poses hypothetical questions\n- Teacher answers and extends\n\nPart Three: Application Practice (about 3 minutes)\n- Teacher gives question\n- Student thinks and answers\n- Teacher provides feedback (affirmation or guidance)\n\nSummary (about 2 minutes)\n- Student attempts to summarize what was learned\n- Teacher supplements and affirms\n- Student expresses gains, connects to practical application\n- Exchange farewells\n```\n\nScript Requirements:\n- Dialogue natural, matches real teacher-student conversation rhythm\n- Avoid written expression\n- Include interjections (\"um\", \"well\", \"oh right\")\n- Allow student to \"interrupt\" with questions\n- All examples sourced from user's interest field\n- About 150-180 words per minute\n\n### Character Settings\n\n**Teacher Character:**\n- Professional yet approachable\n- Good at using metaphors to explain complex concepts\n- Patient in answering questions\n- Timely encouragement and affirmation\n\n**Student Character:**\n- Curious, actively asks questions\n- Represents target user's perspective\n- Makes common mistakes, raises typical confusions\n- Has own interest background (consistent with user settings)\n\n**Step 2: Generate Audio**\n\nUse audio generation tool to convert script to audio:\n- Teacher voice: Warm, professional, patient\n- Student voice: Curious, lively, sincere\n- Speed: Medium for concept explanation, natural rhythm for dialogue, slightly faster for summary\n\n**Step 3: Output**\n- File: `{topic}_audio.mp3`\n- Only return file path, no preview or playback in conversation\n- No auto-play\n- Do not output script files or production notes\n\n### Audio Quality Standards\n1. **Listening Experience** — Sounds like real conversation, not script reading; rhythm varies; key content emphasized\n2. **Learning Effect** — Concept explanation clear; student questions represent real confusion; practice section has testing effect\n3. **Personalization** — Examples 100% from user's interest field; student character gives user identification; language style matches grade\n4. **Audio Quality** — Clear sound; duration about 15 minutes; directly playable\n\n---\n\n## Critical Constraints\n\n1. **Content Fidelity** — All content must be based on original textbook/source material. No unverified information added.\n2. **Grade Adaptation** — Adjust content depth and expression based on grade level for ALL formats.\n3. **Output Rules** — Only return file paths. No inline display of images/PDFs/audio/video. No auto-play. No intermediate files.\n4. **Color Constraints (Notes)** — Maximum 2 colors per PDF. Default blue + pink.\n5. **Color Constraints (Slides)** — Do NOT use blue or purple as theme/background color unless user explicitly requests.\n6. **Image Quality** — Notes: 150 DPI, 90% compression. Slides: 150 DPI, 95% quality.\n7. **Mind Map Depth** — No more than 4 levels. Node text no more than 10 characters.\n8. **Quiz Minimalism** — No UI frameworks, no decorative elements, system default font only.\n\n## Common Mistakes to Avoid\n\n1. **Adding unverified information** — Stick to the source material only\n2. **Ignoring grade level** — Elementary content should not use university-level terminology\n3. **Previewing outputs in conversation** — Never display images, PDFs, or play audio inline\n4. **Dense annotations on notes** — Keep 3-8 annotations per page, not more\n5. **Decorative slides** — Visuals must convey information through structure, not just atmosphere\n6. **Text-heavy slides** — Diagrams should be primary carriers, not text lists with decorative backgrounds\n7. **Using blue/purple in slides** — Forbidden unless user explicitly requests\n8. **Flat quiz feedback** — \"The answer is X\" has no teaching value; always explain why\n9. **Robotic audio dialogue** — Must sound like natural conversation with interjections and interruptions\n10. **Outputting intermediate files** — Only deliver final output file paths\n\n## File & Output Conventions\n\n| Format | Filename Pattern | File Type |\n|--------|-----------------|-----------|\n| Notes | `{topic}_notes.pdf` | PDF |\n| Quiz | `{topic}_quiz.html` | HTML |\n| Slides | `{topic}_slides.pdf`, `{topic}_slides.pptx` | PDF, PPTX |\n| Mind Map | `{topic}_mindmap.png` | PNG |\n| Audio | `{topic}_audio.mp3` | MP3 |\n\nAll files use the topic name as prefix. Deliver all outputs together using `<deliver_assets>` format after all generation is complete.\n\nFile v1.0.5:_meta.json\n\n{\n  \"ownerId\": \"kn712egma3976f4b68hksznyrs85wsr5\",\n  \"slug\": \"eric-knowledge-digest\",\n  \"version\": \"1.0.5\",\n  \"publishedAt\": 1777953417393\n}\n\nArchive v1.0.10: 2 files, 10473 bytes\n\nFiles: _meta.json (141b), SKILL.md (26234b)\n\nFile v1.0.10:SKILL.md\n\n---\nname: knowledge-digest\ndescription: \"Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, and mind maps. Trigger: learning materials, convert textbook, study notes, quiz generation, slides from PDF, mind map, audio course.\"\n---\n\n# KnowledgeDigest — Unified Learning Content Converter\n\n## Overview\n\nKnowledgeDigest converts textbooks, PDFs, or topic descriptions into personalized, multimodal learning experiences. It analyzes source content, then generates any combination of: handwritten-style notes (PDF), interactive quiz webpages (HTML), slides (PDF+PPTX), mind maps (image+Mermaid), and audio courses (MP3). All output is adapted to the learner's grade level and interests.\n\n## Workflow\n\n### Phase 1: Gather User Input\n\n1. Identify what the user has provided:\n   - Uploaded PDF/textbook file (optional)\n   - Topic/direction description\n   - Grade level (elementary / middle school / high school / university / professional)\n   - Expected output format(s)\n\n2. If no PDF/textbook uploaded and no source materials specified (only topic/direction provided):\n   - Ask user:\n     - **Option A:** \"I have materials, uploading now\"\n     - **Option B:** \"No materials, please search and generate courseware about [topic]\"\n   - If user selects B:\n     - Use search tools to collect authoritative materials on the topic\n     - Organize into structured content, generate a basic courseware PDF\n     - Send PDF to user for confirmation: \"This is the basic material I compiled for [topic], please confirm if usable?\"\n     - Continue after user confirmation\n\n3. **Default output formats** (if user does not specify): mindmap + slides (PDF only) + quiz\n\n### Phase 2: Content Analysis\n\nParse the PDF or structured content to extract:\n\n**Document Parsing:**\n- Identify chapter structure (chapters, sections, subsections)\n- Extract heading hierarchy and table of contents\n- Identify body text, images, tables, formulas, and other elements\n\n**Core Concept Extraction:**\n- Identify core concepts and key terms in each chapter\n- Extract definitions, theorems, formulas, and important content\n- Mark difficult points and key knowledge\n\n**Learning Objective Analysis:**\n- Infer learning objectives for each chapter\n- Identify prerequisite knowledge requirements\n- Analyze dependencies between knowledge points\n\n**Output structured analysis results in this format:**\n\n```json\n{\n  \"document_info\": {\n    \"title\": \"Document title\",\n    \"total_pages\": 100,\n    \"language\": \"zh/en\",\n    \"subject\": \"Subject area\"\n  },\n  \"chapters\": [\n    {\n      \"chapter_id\": \"1\",\n      \"title\": \"Chapter title\",\n      \"page_range\": [1, 20],\n      \"sections\": [\n        {\n          \"section_id\": \"1.1\",\n          \"title\": \"Section title\",\n          \"core_concepts\": [\"Concept 1\", \"Concept 2\"],\n          \"key_terms\": [\n            {\"term\": \"Term\", \"definition\": \"Definition\"}\n          ],\n          \"learning_objectives\": [\"Objective 1\", \"Objective 2\"],\n          \"difficulty\": \"easy/medium/hard\",\n          \"prerequisites\": [\"Prerequisite knowledge\"]\n        }\n      ]\n    }\n  ],\n  \"knowledge_graph\": {\n    \"nodes\": [\"Concept node list\"],\n    \"edges\": [{\"from\": \"Concept A\", \"to\": \"Concept B\", \"relation\": \"depends/contains/related\"}]\n  }\n}\n```\n\n**Parsing Rules:**\n1. Chapter Recognition — Identify hierarchy based on font size, bold, numbering, etc. Handle documents without clear chapter markers by logically segmenting.\n2. Concept Extraction — Identify bolded, highlighted, boxed important content. Extract proper nouns and term definitions. Identify formulas and theorems.\n3. Difficulty Assessment — Assess based on concept abstraction level, prerequisite knowledge, and content complexity.\n4. Quality Assurance — Ensure all chapters identified, verify knowledge point coverage completeness, check accuracy of concept definitions.\n\n### Phase 3: Generate Requested Formats\n\nBased on user-selected output formats, generate each in sequence. For each format, follow the corresponding section below.\n\n### Phase 4: Deliver Assets\n\nAfter all generation is complete:\n- Only return file paths, no previews allowed\n- No inline display of images/PDFs/audio/video in conversation\n- Audio/video files must not auto-play\n\nPresent to user using deliver_assets format:\n```\n<deliver_assets>\n<item>\n<path>file path</path>\n</item>\n</deliver_assets>\n```\n\n---\n\n## Supported Output Formats\n\n| Format | Output | Description |\n|--------|--------|-------------|\n| `notes` | `{topic}_notes.pdf` | Handwritten-style notes (annotated on original or generated from scratch) |\n| `quiz` | `{topic}_quiz.html` | Minimalist interactive HTML quiz with instant feedback |\n| `slides` | `{topic}_slides.pdf` + `{topic}_slides.pptx` | Visual slides |\n| `mindmap` | `{topic}_mindmap.png` + Mermaid text | Mind map image |\n| `audio` | `{topic}_audio.mp3` | Audio course in teacher-student dialogue format |\n| `all` | All of the above | Generate every format |\n\n---\n\n## Personalization: Grade Level Adaptation\n\nAll generated content must be adapted to the learner's grade level:\n\n| Grade | Language & Tone | Content Density | Visual Style |\n|-------|----------------|-----------------|--------------|\n| **Elementary** | Lively, simple Q&A, encouraging, story-style | Low density, more drawings, large font | Fun elements, bright colors, short text |\n| **Middle school** | Guided questioning, moderate challenges, youth-oriented | Moderate, image-text combination, clear labels | Image-text combination, moderate information |\n| **High school** | In-depth discussion, logical reasoning, appropriate academic tone | Higher density, logic diagrams | Professional feel, data visualization |\n| **University/Professional** | Seminar-style, critical thinking, professional terminology | High density, professional charts, complex structures | Academic style, comprehensive application |\n\n**Interest Adaptation** (applies to all formats):\n- Examples and metaphors use the user's interest field\n- Scenarios drawn from the user's familiar domain\n- Visual style and analogies match user interests\n\n---\n\n## Format 1: Notes Generation\n\n### Input Type Determination\n\n**Type A — Existing Paper/Courseware:**\n- PDF format academic papers, courseware/PPT exports, scanned textbook pages\n- Features: Fixed layout, page numbers, chapter numbering, formulas/charts\n- Action: Overlay handwritten notes on original pages\n\n**Type B — Non-existing Content:**\n- Plain text notes, knowledge point lists, oral transcripts, web content excerpts\n- Features: No fixed layout, needs reorganization\n- Action: Generate notes PDF from scratch\n\n### Type A Workflow: Adding Notes to Original Document\n\n**Step 1: Analyze Original Structure**\n\nAnalyze PDF content page by page:\n- Identify chapter titles and positions\n- Identify core concepts/terms\n- Identify formulas and their meanings\n- Identify problem/challenge statements\n- Identify solutions/methods\n- Identify key conclusions\n\n**Step 2: Plan Note Content**\n\nPlan handwritten annotations for each page (3-8 annotations per page, not too dense):\n\nAnnotation Types:\n1. **Chapter title translation/explanation** — e.g., original \"3.1 Preliminaries\" → annotate \"Background Knowledge\"\n2. **Key questions** — e.g., \"Key: How to reduce complexity?\"\n3. **Concept explanation** — e.g., annotate \"kernel trick\" next to formula\n4. **Problem marking** — e.g., \"Problem: memory overflow\"\n5. **Solutions** — e.g., \"Solution: forget gate\"\n6. **Formula notes** — e.g., \"recursive form\", \"write operation & read operation\"\n7. **Structure annotation** — e.g., use braces to mark formula groups, write \"→ O(N²) complexity\" beside\n\nAnnotation Planning Principles:\n- Positions avoid blocking key content\n- Utilize margins and paragraph gaps\n- Related content connected with lines or arrows\n\n**Step 3: Generate Annotated Images**\n\nConvert each PDF page to image, then use image generation tool to add handwritten-style annotations.\n\nHandwritten Annotation Style Requirements:\n- **Font:** Handwritten style, slightly tilted\n- **Color:** Unified colors throughout PDF, no more than 2\n  - Default: blue and pink (unless user specifies otherwise)\n  - All subsequent pages can only choose from these 2 colors\n  - Color assignment rules:\n    - Color 1 (blue/primary): Chapter titles, structure annotations, concept explanations, formula notes\n    - Color 2 (pink/accent): Key questions, problem marking, solutions\n- **Size:** Slightly larger than body text, eye-catching but not overwhelming\n- **Position:** Margins, paragraph gaps, blank space next to formulas\n\n**Step 4: Compile PDF**\n- Maintain original page order\n- Image quality: 150 DPI\n- Compression quality: 90%\n\n### Type B Workflow: Generating Notes from Scratch\n\n**Step 1: Organize Content Structure**\n- Main title → Chapters/modules → Core concepts → Key points/details → Examples/applications\n\n**Step 2: Design Note Layout**\n\nLayout Elements:\n- Title area: Large handwritten title\n- Body area: Handwritten-style bullet points\n- Diagram area: Concept maps, flowcharts, relationship diagrams (hand-drawn style)\n- Annotation area: Key markers, question marks, exclamation marks\n- Blank area: Space reserved for user's own notes\n\n**Step 3: Generate Note Page Images**\n\nEach page contains:\n- Page title (handwritten large text)\n- Core content (handwritten bullet points)\n- Diagrams (hand-drawn style concept maps/flowcharts)\n- Key annotations (boxes, arrows, underlines)\n- Notes (like \"Important!\", \"Common mistake\", \"Remember this\")\n\nStyle Requirements:\n- **Overall:** Looks like carefully made student notes, not printed document\n- **Font:** Handwritten, varying sizes (large for titles, medium for body, small for notes)\n- **Color:** Unified colors throughout PDF, no more than 2\n  - Default: blue and pink (unless user specifies otherwise)\n  - Color assignment: Blue (titles, framework, notes), Pink (key points)\n- **Layout:** Organized but not rigid, slight tilting and variation allowed\n- **Elements:** Arrows, underlines, boxes, cloud frames, asterisks — use only when necessary\n\n**Step 4: Compile PDF**\n- Arrange in logical content order\n- Image quality: 150 DPI, compression quality: 90%\n\n### Notes Output\n- File: `{topic}_notes.pdf`\n- Only return file path, no preview in conversation\n- Do not output intermediate image files or content scripts\n\n### Notes Quality Standards\n1. **Content Accuracy** — Annotations based on original text; translation/explanation accurate; no added information\n2. **Annotation Value** — Annotations help understanding, not simple repetition; key points highlight important concepts; problems and solutions correspond clearly\n3. **Visual Effect** — Handwritten style natural, not machine-printed; color coordination harmonious; annotation positions reasonable\n4. **Usability** — PDF printable; suitable for screen reading; reasonable file size\n\n---\n\n## Format 2: Quiz Generation\n\n### Question Design\n\nAt least 5 questions per section. Distribution:\n- Multiple choice (multiple_choice): 2-3 questions\n- True/false (true_false): 1-2 questions\n- Fill in the blank (fill_blank): 1-2 questions\n\nDifficulty distribution:\n- 40% Easy (memory, comprehension)\n- 40% Medium (application)\n- 20% Hard (analysis, synthesis)\n\nEach question must include:\n- Question content (using personalized scenario)\n- Correct answer\n- Answer explanation (has teaching value, not just \"the answer is X\")\n- Related core concept\n\n### HTML Generation\n\nGenerate a single HTML file containing all questions and interaction logic.\n\n**Design Principle: Minimalist**\n\nVisual Style:\n- Pure white background\n- Black text\n- No decorative elements, no icons, no gradients, no shadows\n- No borders or only 1px gray thin lines\n- Font: System default font\n- Minimal CSS, no UI frameworks\n\nInteraction Design:\n- Click option to select, selected state distinguished by slight background color\n- Show correct/incorrect and explanation immediately after submit\n- Correct: Green text \"Correct\"\n- Incorrect: Red text \"Incorrect\" + correct answer + explanation\n- Show total score at end\n\n**HTML Structure Template:**\n\n```html\n<!DOCTYPE html>\n<html lang=\"en\">\n<head>\n  <meta charset=\"UTF-8\">\n  <meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">\n  <title>Chapter Quiz</title>\n  <style>\n    body {\n      font-family: system-ui, sans-serif;\n      max-width: 600px;\n      margin: 40px auto;\n      padding: 20px;\n      line-height: 1.6;\n    }\n    h1 { font-size: 1.5em; font-weight: normal; }\n    .question { margin: 30px 0; }\n    .question-text { margin-bottom: 15px; }\n    .option {\n      display: block;\n      padding: 10px;\n      margin: 5px 0;\n      cursor: pointer;\n    }\n    .option:hover { background: #f5f5f5; }\n    .option.selected { background: #e8e8e8; }\n    .feedback { margin-top: 10px; font-size: 0.9em; }\n    .correct { color: #2e7d32; }\n    .incorrect { color: #c62828; }\n    .explanation { color: #666; margin-top: 5px; }\n    button {\n      padding: 10px 20px;\n      background: #333;\n      color: white;\n      border: none;\n      cursor: pointer;\n      margin-top: 20px;\n    }\n    .score { font-size: 1.2em; margin-top: 30px; }\n  </style>\n</head>\n<body>\n  <h1>Chapter Title - Quiz</h1>\n\n  <div class=\"question\" data-answer=\"A\">\n    <div class=\"question-text\">1. Question content</div>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"A\"> A. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"B\"> B. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"C\"> C. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"D\"> D. Option</label>\n    <div class=\"feedback\"></div>\n  </div>\n\n  <!-- More questions... -->\n\n  <button onclick=\"submit()\">Submit</button>\n  <div class=\"score\"></div>\n\n  <script>\n    const explanations = {\n      q1: \"Explanation content...\",\n      // ...\n    };\n\n    function submit() {\n      let correct = 0;\n      document.querySelectorAll('.question').forEach((q, i) => {\n        const answer = q.dataset.answer;\n        const selected = q.querySelector('input:checked');\n        const feedback = q.querySelector('.feedback');\n        const qName = 'q' + (i + 1);\n\n        if (selected && selected.value === answer) {\n          feedback.innerHTML = '<span class=\"correct\">Correct</span>';\n          correct++;\n        } else {\n          feedback.innerHTML = '<span class=\"incorrect\">Incorrect</span> Correct answer: ' + answer +\n            '<div class=\"explanation\">' + explanations[qName] + '</div>';\n        }\n      });\n\n      document.querySelector('.score').textContent =\n        'Score: ' + correct + '/' + document.querySelectorAll('.question').length;\n    }\n  </script>\n</body>\n</html>\n```\n\n### Quiz Output\n- File: `{topic}_quiz.html`\n- Only return file path, no preview in conversation\n- Do not output JSON data, CSS files, or JS files separately\n\n### Quiz Quality Standards\n1. **Content Accuracy** — All knowledge points based on original textbook; answers and explanations correct; question wording clear and unambiguous\n2. **Personalization** — Question scenarios match user interests; difficulty matches grade level; language style suits target audience\n3. **Interaction Experience** — Click response instant; feedback clear; explanations have teaching value\n4. **Visual Minimalism** — No decorative elements; no framework dependencies; file size minimized\n\n---\n\n## Format 3: Slides Generation\n\n### Design Considerations\n\nTreat these as a flexible menu, not a mandatory checklist:\n\n1. **Topic, Purpose & Audience** — What is this about? Who needs to understand it? Where will it be presented?\n2. **Content Foundation & Sources** — What materials or data need to be presented?\n3. **Visual Approach (CRITICAL)**\n   - Default to explanatory visuals: cutaway views, annotated structure diagrams, exploded views, schematic illustrations\n   - Visual elements are primary information carriers, not decorative backgrounds for text lists\n   - Default information density matches professional infographics and technical illustrations\n   - **CRITICAL:** Diagrams must convey information through structure, not just provide atmosphere. Text should be labels/annotations, not main content. Reject purely decorative visuals with core information dependent on text lists\n   - Reject the inefficient pattern of \"large white space + centered single line of text\"\n4. **Narrative Flow & Chapters** — How should viewers move through the content? How is slide flow arranged?\n5. **Text Style & Density**\n   - Language: Explanatory text uses language explicitly requested by user, otherwise match user's conversation language\n   - Typography: Chinese and English titles preferably use serif fonts (Chinese uses Song font family)\n6. **Visual Style, Color & Mood**\n   - Visual language of encyclopedias and reference books: explanatory diagrams, cutaway illustrations, annotated structures\n   - Refined spatial composition and typographic precision of high-end journals\n   - Intentional asymmetry and layered information design of contemporary design publications\n   - Apply asymmetric grids, intentional breathing space, layered information organization, diagonal composition, dynamic typography as internalized design language\n   - **Color restriction:** Unless user explicitly specifies, do NOT use blue or purple as theme color or background color\n\n### Slides Workflow\n\n**Step 1: Design Strategy — Create Content Script**\n\nInformation architecture first: Structure content into hierarchical slides, each slide as an information unit defined by what data/facts/relationships it carries. Let content volume naturally determine slide count.\n\nOutput `content_script.md`:\n\n```\n# Slides Content Script\n\n## Slide 1: [Title]\n**Subtopic A**: [Label]\n[50-80 word narrative paragraph describing information content to be visualized]\n\n**Subtopic B**: [Label]\n[50-80 word narrative paragraph]\n\n## Slide 2: [Title]\n...\n```\n\nContent Script Specification:\n- Only describe \"what information needs to be presented\", not \"how to present it\"\n- Do NOT include \"Visual Description\" sections\n- Do NOT describe colors, backgrounds, decorative elements, atmosphere effects, mood, or layout details\n- Focus on pure information architecture\n- 2-3 focused subtopics per slide\n\n**Step 2: Sequential Image Generation**\n\nUse image generation tool to generate slides one by one:\n- First slide: Use gen_images (create from scratch)\n- Subsequent slides: Use edit_images, base_image_file points to previous slide\n\nFormat: Default 16:9 landscape ratio. Save each slide image locally.\n\n**Prompt Construction for Each Slide — Must include these 6 points:**\n\n1. **Visualization Type** — Prioritize diagram forms over text-dominated presentations: cutaway views, flowcharts, annotated structure diagrams, relationship diagrams, timeline overlays. Integrate multiple subtopics into unified visual structure. Avoid \"parallel cards/grid displays/multi-column layouts\" and text-heavy traditional typography.\n\n2. **Information Hierarchy** — Primary and secondary information distinguished through visual hierarchy (size, position, contrast). Not flat lists.\n\n3. **Composition Instructions** — Asymmetric layout, diagonal momentum, and other methods to break rigid symmetry.\n\n4. **Density Requirements** — Clear information hierarchy over quantity. Appropriate white space serves readability, but not empty and sparse.\n\n5. **Layout Independence** — Explicitly state this slide's visualization type is chosen based on its content, not copying previous slide. Re-evaluate what this specific content needs. But describe inherited elements in detail.\n\n6. **Style Consistency** — If user provided visual style or reference images, each prompt must describe that style's characteristics in detail.\n\n**Step 3: Compile Output**\n\nAfter generating all slide images:\n- Auto-compile into PDF (150 DPI, 95% quality, controlled file size)\n- Auto-compile into PPTX presentation\n\n### Slides Output\n- Files: `{topic}_slides.pdf` + `{topic}_slides.pptx`\n- Only return file paths, no preview in conversation\n- Do not output individual slide images, summary documents, content outlines, design descriptions, or usage instructions\n\n---\n\n## Format 4: Mind Map Generation\n\n### Mind Map Workflow\n\n**Step 1: Design Content Structure**\n\nDetermine node hierarchy and relationships:\n- Root node: Chapter theme\n- Level 1 nodes: Core concepts\n- Level 2 nodes: Detail points\n- No more than 4 levels\n- Each node text concise (no more than 10 characters)\n- Mark relationships between concepts (parallel/progressive/causal/contrast)\n\n**Step 2: Generate Image**\n\nUse gen_images to generate mind map image:\n- Format: 16:9 or square (based on content)\n- Style: Clear visual hierarchy, professional infographic style\n\n**Step 3: Output**\n\n- Mind map image: `{topic}_mindmap.png`\n- Attached Mermaid format text (optional, for users who need to edit)\n- Only return file path, no image preview in conversation\n\n---\n\n## Format 5: Audio Course Generation\n\n### Audio Workflow\n\n**Step 1: Write Dialogue Script**\n\nWrite teacher-student dialogue script:\n\n```\nOpening (about 1 minute)\n- Teacher greets, introduces today's topic\n- Student responds, expresses existing knowledge or questions\n- Teacher builds connection using user's interest field\n\nPart One: Concept Introduction (about 4 minutes)\n- Teacher asks questions from user's interest scenario\n- Student observes/answers\n- Teacher introduces core concept, defines in conversational manner\n- Student requests examples\n- Teacher explains in detail with personalized examples\n- Student restates in own words to confirm understanding\n\nPart Two: Deep Understanding (about 5 minutes)\n- Teacher explains important characteristics of concept\n- Student raises common confusion/misconception\n- Teacher clarifies misconception\n- Student poses hypothetical questions\n- Teacher answers and extends\n\nPart Three: Application Practice (about 3 minutes)\n- Teacher gives question\n- Student thinks and answers\n- Teacher provides feedback (affirmation or guidance)\n\nSummary (about 2 minutes)\n- Student attempts to summarize what was learned\n- Teacher supplements and affirms\n- Student expresses gains, connects to practical application\n- Exchange farewells\n```\n\nScript Requirements:\n- Dialogue natural, matches real teacher-student conversation rhythm\n- Avoid written expression\n- Include interjections (\"um\", \"well\", \"oh right\")\n- Allow student to \"interrupt\" with questions\n- All examples sourced from user's interest field\n- About 150-180 words per minute\n\n### Character Settings\n\n**Teacher Character:**\n- Professional yet approachable\n- Good at using metaphors to explain complex concepts\n- Patient in answering questions\n- Timely encouragement and affirmation\n\n**Student Character:**\n- Curious, actively asks questions\n- Represents target user's perspective\n- Makes common mistakes, raises typical confusions\n- Has own interest background (consistent with user settings)\n\n**Step 2: Generate Audio**\n\nUse audio generation tool to convert script to audio:\n- Teacher voice: Warm, professional, patient\n- Student voice: Curious, lively, sincere\n- Speed: Medium for concept explanation, natural rhythm for dialogue, slightly faster for summary\n\n**Step 3: Output**\n- File: `{topic}_audio.mp3`\n- Only return file path, no preview or playback in conversation\n- No auto-play\n- Do not output script files or production notes\n\n### Audio Quality Standards\n1. **Listening Experience** — Sounds like real conversation, not script reading; rhythm varies; key content emphasized\n2. **Learning Effect** — Concept explanation clear; student questions represent real confusion; practice section has testing effect\n3. **Personalization** — Examples 100% from user's interest field; student character gives user identification; language style matches grade\n4. **Audio Quality** — Clear sound; duration about 15 minutes; directly playable\n\n---\n\n## Critical Constraints\n\n1. **Content Fidelity** — All content must be based on original textbook/source material. No unverified information added.\n2. **Grade Adaptation** — Adjust content depth and expression based on grade level for ALL formats.\n3. **Output Rules** — Only return file paths. No inline display of images/PDFs/audio/video. No auto-play. No intermediate files.\n4. **Color Constraints (Notes)** — Maximum 2 colors per PDF. Default blue + pink.\n5. **Color Constraints (Slides)** — Do NOT use blue or purple as theme/background color unless user explicitly requests.\n6. **Image Quality** — Notes: 150 DPI, 90% compression. Slides: 150 DPI, 95% quality.\n7. **Mind Map Depth** — No more than 4 levels. Node text no more than 10 characters.\n8. **Quiz Minimalism** — No UI frameworks, no decorative elements, system default font only.\n\n## Common Mistakes to Avoid\n\n1. **Adding unverified information** — Stick to the source material only\n2. **Ignoring grade level** — Elementary content should not use university-level terminology\n3. **Previewing outputs in conversation** — Never display images, PDFs, or play audio inline\n4. **Dense annotations on notes** — Keep 3-8 annotations per page, not more\n5. **Decorative slides** — Visuals must convey information through structure, not just atmosphere\n6. **Text-heavy slides** — Diagrams should be primary carriers, not text lists with decorative backgrounds\n7. **Using blue/purple in slides** — Forbidden unless user explicitly requests\n8. **Flat quiz feedback** — \"The answer is X\" has no teaching value; always explain why\n9. **Robotic audio dialogue** — Must sound like natural conversation with interjections and interruptions\n10. **Outputting intermediate files** — Only deliver final output file paths\n\n## File & Output Conventions\n\n| Format | Filename Pattern | File Type |\n|--------|-----------------|-----------|\n| Notes | `{topic}_notes.pdf` | PDF |\n| Quiz | `{topic}_quiz.html` | HTML |\n| Slides | `{topic}_slides.pdf`, `{topic}_slides.pptx` | PDF, PPTX |\n| Mind Map | `{topic}_mindmap.png` | PNG |\n| Audio | `{topic}_audio.mp3` | MP3 |\n\nAll files use the topic name as prefix. Deliver all outputs together using `<deliver_assets>` format after all generation is complete.\n\nFile v1.0.10:_meta.json\n\n{\n  \"ownerId\": \"kn712egma3976f4b68hksznyrs85wsr5\",\n  \"slug\": \"eric-knowledge-digest\",\n  \"version\": \"1.0.10\",\n  \"publishedAt\": 1777917395105\n}\n\nArchive v11.0.0: 3 files, 11593 bytes\n\nFiles: _meta.json (141b), skill-card.md (1977b), SKILL.md (26234b)\n\nFile v11.0.0:SKILL.md\n\n---\nname: knowledge-digest\ndescription: \"Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, and mind maps. Trigger: learning materials, convert textbook, study notes, quiz generation, slides from PDF, mind map, audio course.\"\n---\n\n# KnowledgeDigest — Unified Learning Content Converter\n\n## Overview\n\nKnowledgeDigest converts textbooks, PDFs, or topic descriptions into personalized, multimodal learning experiences. It analyzes source content, then generates any combination of: handwritten-style notes (PDF), interactive quiz webpages (HTML), slides (PDF+PPTX), mind maps (image+Mermaid), and audio courses (MP3). All output is adapted to the learner's grade level and interests.\n\n## Workflow\n\n### Phase 1: Gather User Input\n\n1. Identify what the user has provided:\n   - Uploaded PDF/textbook file (optional)\n   - Topic/direction description\n   - Grade level (elementary / middle school / high school / university / professional)\n   - Expected output format(s)\n\n2. If no PDF/textbook uploaded and no source materials specified (only topic/direction provided):\n   - Ask user:\n     - **Option A:** \"I have materials, uploading now\"\n     - **Option B:** \"No materials, please search and generate courseware about [topic]\"\n   - If user selects B:\n     - Use search tools to collect authoritative materials on the topic\n     - Organize into structured content, generate a basic courseware PDF\n     - Send PDF to user for confirmation: \"This is the basic material I compiled for [topic], please confirm if usable?\"\n     - Continue after user confirmation\n\n3. **Default output formats** (if user does not specify): mindmap + slides (PDF only) + quiz\n\n### Phase 2: Content Analysis\n\nParse the PDF or structured content to extract:\n\n**Document Parsing:**\n- Identify chapter structure (chapters, sections, subsections)\n- Extract heading hierarchy and table of contents\n- Identify body text, images, tables, formulas, and other elements\n\n**Core Concept Extraction:**\n- Identify core concepts and key terms in each chapter\n- Extract definitions, theorems, formulas, and important content\n- Mark difficult points and key knowledge\n\n**Learning Objective Analysis:**\n- Infer learning objectives for each chapter\n- Identify prerequisite knowledge requirements\n- Analyze dependencies between knowledge points\n\n**Output structured analysis results in this format:**\n\n```json\n{\n  \"document_info\": {\n    \"title\": \"Document title\",\n    \"total_pages\": 100,\n    \"language\": \"zh/en\",\n    \"subject\": \"Subject area\"\n  },\n  \"chapters\": [\n    {\n      \"chapter_id\": \"1\",\n      \"title\": \"Chapter title\",\n      \"page_range\": [1, 20],\n      \"sections\": [\n        {\n          \"section_id\": \"1.1\",\n          \"title\": \"Section title\",\n          \"core_concepts\": [\"Concept 1\", \"Concept 2\"],\n          \"key_terms\": [\n            {\"term\": \"Term\", \"definition\": \"Definition\"}\n          ],\n          \"learning_objectives\": [\"Objective 1\", \"Objective 2\"],\n          \"difficulty\": \"easy/medium/hard\",\n          \"prerequisites\": [\"Prerequisite knowledge\"]\n        }\n      ]\n    }\n  ],\n  \"knowledge_graph\": {\n    \"nodes\": [\"Concept node list\"],\n    \"edges\": [{\"from\": \"Concept A\", \"to\": \"Concept B\", \"relation\": \"depends/contains/related\"}]\n  }\n}\n```\n\n**Parsing Rules:**\n1. Chapter Recognition — Identify hierarchy based on font size, bold, numbering, etc. Handle documents without clear chapter markers by logically segmenting.\n2. Concept Extraction — Identify bolded, highlighted, boxed important content. Extract proper nouns and term definitions. Identify formulas and theorems.\n3. Difficulty Assessment — Assess based on concept abstraction level, prerequisite knowledge, and content complexity.\n4. Quality Assurance — Ensure all chapters identified, verify knowledge point coverage completeness, check accuracy of concept definitions.\n\n### Phase 3: Generate Requested Formats\n\nBased on user-selected output formats, generate each in sequence. For each format, follow the corresponding section below.\n\n### Phase 4: Deliver Assets\n\nAfter all generation is complete:\n- Only return file paths, no previews allowed\n- No inline display of images/PDFs/audio/video in conversation\n- Audio/video files must not auto-play\n\nPresent to user using deliver_assets format:\n```\n<deliver_assets>\n<item>\n<path>file path</path>\n</item>\n</deliver_assets>\n```\n\n---\n\n## Supported Output Formats\n\n| Format | Output | Description |\n|--------|--------|-------------|\n| `notes` | `{topic}_notes.pdf` | Handwritten-style notes (annotated on original or generated from scratch) |\n| `quiz` | `{topic}_quiz.html` | Minimalist interactive HTML quiz with instant feedback |\n| `slides` | `{topic}_slides.pdf` + `{topic}_slides.pptx` | Visual slides |\n| `mindmap` | `{topic}_mindmap.png` + Mermaid text | Mind map image |\n| `audio` | `{topic}_audio.mp3` | Audio course in teacher-student dialogue format |\n| `all` | All of the above | Generate every format |\n\n---\n\n## Personalization: Grade Level Adaptation\n\nAll generated content must be adapted to the learner's grade level:\n\n| Grade | Language & Tone | Content Density | Visual Style |\n|-------|----------------|-----------------|--------------|\n| **Elementary** | Lively, simple Q&A, encouraging, story-style | Low density, more drawings, large font | Fun elements, bright colors, short text |\n| **Middle school** | Guided questioning, moderate challenges, youth-oriented | Moderate, image-text combination, clear labels | Image-text combination, moderate information |\n| **High school** | In-depth discussion, logical reasoning, appropriate academic tone | Higher density, logic diagrams | Professional feel, data visualization |\n| **University/Professional** | Seminar-style, critical thinking, professional terminology | High density, professional charts, complex structures | Academic style, comprehensive application |\n\n**Interest Adaptation** (applies to all formats):\n- Examples and metaphors use the user's interest field\n- Scenarios drawn from the user's familiar domain\n- Visual style and analogies match user interests\n\n---\n\n## Format 1: Notes Generation\n\n### Input Type Determination\n\n**Type A — Existing Paper/Courseware:**\n- PDF format academic papers, courseware/PPT exports, scanned textbook pages\n- Features: Fixed layout, page numbers, chapter numbering, formulas/charts\n- Action: Overlay handwritten notes on original pages\n\n**Type B — Non-existing Content:**\n- Plain text notes, knowledge point lists, oral transcripts, web content excerpts\n- Features: No fixed layout, needs reorganization\n- Action: Generate notes PDF from scratch\n\n### Type A Workflow: Adding Notes to Original Document\n\n**Step 1: Analyze Original Structure**\n\nAnalyze PDF content page by page:\n- Identify chapter titles and positions\n- Identify core concepts/terms\n- Identify formulas and their meanings\n- Identify problem/challenge statements\n- Identify solutions/methods\n- Identify key conclusions\n\n**Step 2: Plan Note Content**\n\nPlan handwritten annotations for each page (3-8 annotations per page, not too dense):\n\nAnnotation Types:\n1. **Chapter title translation/explanation** — e.g., original \"3.1 Preliminaries\" → annotate \"Background Knowledge\"\n2. **Key questions** — e.g., \"Key: How to reduce complexity?\"\n3. **Concept explanation** — e.g., annotate \"kernel trick\" next to formula\n4. **Problem marking** — e.g., \"Problem: memory overflow\"\n5. **Solutions** — e.g., \"Solution: forget gate\"\n6. **Formula notes** — e.g., \"recursive form\", \"write operation & read operation\"\n7. **Structure annotation** — e.g., use braces to mark formula groups, write \"→ O(N²) complexity\" beside\n\nAnnotation Planning Principles:\n- Positions avoid blocking key content\n- Utilize margins and paragraph gaps\n- Related content connected with lines or arrows\n\n**Step 3: Generate Annotated Images**\n\nConvert each PDF page to image, then use image generation tool to add handwritten-style annotations.\n\nHandwritten Annotation Style Requirements:\n- **Font:** Handwritten style, slightly tilted\n- **Color:** Unified colors throughout PDF, no more than 2\n  - Default: blue and pink (unless user specifies otherwise)\n  - All subsequent pages can only choose from these 2 colors\n  - Color assignment rules:\n    - Color 1 (blue/primary): Chapter titles, structure annotations, concept explanations, formula notes\n    - Color 2 (pink/accent): Key questions, problem marking, solutions\n- **Size:** Slightly larger than body text, eye-catching but not overwhelming\n- **Position:** Margins, paragraph gaps, blank space next to formulas\n\n**Step 4: Compile PDF**\n- Maintain original page order\n- Image quality: 150 DPI\n- Compression quality: 90%\n\n### Type B Workflow: Generating Notes from Scratch\n\n**Step 1: Organize Content Structure**\n- Main title → Chapters/modules → Core concepts → Key points/details → Examples/applications\n\n**Step 2: Design Note Layout**\n\nLayout Elements:\n- Title area: Large handwritten title\n- Body area: Handwritten-style bullet points\n- Diagram area: Concept maps, flowcharts, relationship diagrams (hand-drawn style)\n- Annotation area: Key markers, question marks, exclamation marks\n- Blank area: Space reserved for user's own notes\n\n**Step 3: Generate Note Page Images**\n\nEach page contains:\n- Page title (handwritten large text)\n- Core content (handwritten bullet points)\n- Diagrams (hand-drawn style concept maps/flowcharts)\n- Key annotations (boxes, arrows, underlines)\n- Notes (like \"Important!\", \"Common mistake\", \"Remember this\")\n\nStyle Requirements:\n- **Overall:** Looks like carefully made student notes, not printed document\n- **Font:** Handwritten, varying sizes (large for titles, medium for body, small for notes)\n- **Color:** Unified colors throughout PDF, no more than 2\n  - Default: blue and pink (unless user specifies otherwise)\n  - Color assignment: Blue (titles, framework, notes), Pink (key points)\n- **Layout:** Organized but not rigid, slight tilting and variation allowed\n- **Elements:** Arrows, underlines, boxes, cloud frames, asterisks — use only when necessary\n\n**Step 4: Compile PDF**\n- Arrange in logical content order\n- Image quality: 150 DPI, compression quality: 90%\n\n### Notes Output\n- File: `{topic}_notes.pdf`\n- Only return file path, no preview in conversation\n- Do not output intermediate image files or content scripts\n\n### Notes Quality Standards\n1. **Content Accuracy** — Annotations based on original text; translation/explanation accurate; no added information\n2. **Annotation Value** — Annotations help understanding, not simple repetition; key points highlight important concepts; problems and solutions correspond clearly\n3. **Visual Effect** — Handwritten style natural, not machine-printed; color coordination harmonious; annotation positions reasonable\n4. **Usability** — PDF printable; suitable for screen reading; reasonable file size\n\n---\n\n## Format 2: Quiz Generation\n\n### Question Design\n\nAt least 5 questions per section. Distribution:\n- Multiple choice (multiple_choice): 2-3 questions\n- True/false (true_false): 1-2 questions\n- Fill in the blank (fill_blank): 1-2 questions\n\nDifficulty distribution:\n- 40% Easy (memory, comprehension)\n- 40% Medium (application)\n- 20% Hard (analysis, synthesis)\n\nEach question must include:\n- Question content (using personalized scenario)\n- Correct answer\n- Answer explanation (has teaching value, not just \"the answer is X\")\n- Related core concept\n\n### HTML Generation\n\nGenerate a single HTML file containing all questions and interaction logic.\n\n**Design Principle: Minimalist**\n\nVisual Style:\n- Pure white background\n- Black text\n- No decorative elements, no icons, no gradients, no shadows\n- No borders or only 1px gray thin lines\n- Font: System default font\n- Minimal CSS, no UI frameworks\n\nInteraction Design:\n- Click option to select, selected state distinguished by slight background color\n- Show correct/incorrect and explanation immediately after submit\n- Correct: Green text \"Correct\"\n- Incorrect: Red text \"Incorrect\" + correct answer + explanation\n- Show total score at end\n\n**HTML Structure Template:**\n\n```html\n<!DOCTYPE html>\n<html lang=\"en\">\n<head>\n  <meta charset=\"UTF-8\">\n  <meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">\n  <title>Chapter Quiz</title>\n  <style>\n    body {\n      font-family: system-ui, sans-serif;\n      max-width: 600px;\n      margin: 40px auto;\n      padding: 20px;\n      line-height: 1.6;\n    }\n    h1 { font-size: 1.5em; font-weight: normal; }\n    .question { margin: 30px 0; }\n    .question-text { margin-bottom: 15px; }\n    .option {\n      display: block;\n      padding: 10px;\n      margin: 5px 0;\n      cursor: pointer;\n    }\n    .option:hover { background: #f5f5f5; }\n    .option.selected { background: #e8e8e8; }\n    .feedback { margin-top: 10px; font-size: 0.9em; }\n    .correct { color: #2e7d32; }\n    .incorrect { color: #c62828; }\n    .explanation { color: #666; margin-top: 5px; }\n    button {\n      padding: 10px 20px;\n      background: #333;\n      color: white;\n      border: none;\n      cursor: pointer;\n      margin-top: 20px;\n    }\n    .score { font-size: 1.2em; margin-top: 30px; }\n  </style>\n</head>\n<body>\n  <h1>Chapter Title - Quiz</h1>\n\n  <div class=\"question\" data-answer=\"A\">\n    <div class=\"question-text\">1. Question content</div>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"A\"> A. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"B\"> B. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"C\"> C. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"D\"> D. Option</label>\n    <div class=\"feedback\"></div>\n  </div>\n\n  <!-- More questions... -->\n\n  <button onclick=\"submit()\">Submit</button>\n  <div class=\"score\"></div>\n\n  <script>\n    const explanations = {\n      q1: \"Explanation content...\",\n      // ...\n    };\n\n    function submit() {\n      let correct = 0;\n      document.querySelectorAll('.question').forEach((q, i) => {\n        const answer = q.dataset.answer;\n        const selected = q.querySelector('input:checked');\n        const feedback = q.querySelector('.feedback');\n        const qName = 'q' + (i + 1);\n\n        if (selected && selected.value === answer) {\n          feedback.innerHTML = '<span class=\"correct\">Correct</span>';\n          correct++;\n        } else {\n          feedback.innerHTML = '<span class=\"incorrect\">Incorrect</span> Correct answer: ' + answer +\n            '<div class=\"explanation\">' + explanations[qName] + '</div>';\n        }\n      });\n\n      document.querySelector('.score').textContent =\n        'Score: ' + correct + '/' + document.querySelectorAll('.question').length;\n    }\n  </script>\n</body>\n</html>\n```\n\n### Quiz Output\n- File: `{topic}_quiz.html`\n- Only return file path, no preview in conversation\n- Do not output JSON data, CSS files, or JS files separately\n\n### Quiz Quality Standards\n1. **Content Accuracy** — All knowledge points based on original textbook; answers and explanations correct; question wording clear and unambiguous\n2. **Personalization** — Question scenarios match user interests; difficulty matches grade level; language style suits target audience\n3. **Interaction Experience** — Click response instant; feedback clear; explanations have teaching value\n4. **Visual Minimalism** — No decorative elements; no framework dependencies; file size minimized\n\n---\n\n## Format 3: Slides Generation\n\n### Design Considerations\n\nTreat these as a flexible menu, not a mandatory checklist:\n\n1. **Topic, Purpose & Audience** — What is this about? Who needs to understand it? Where will it be presented?\n2. **Content Foundation & Sources** — What materials or data need to be presented?\n3. **Visual Approach (CRITICAL)**\n   - Default to explanatory visuals: cutaway views, annotated structure diagrams, exploded views, schematic illustrations\n   - Visual elements are primary information carriers, not decorative backgrounds for text lists\n   - Default information density matches professional infographics and technical illustrations\n   - **CRITICAL:** Diagrams must convey information through structure, not just provide atmosphere. Text should be labels/annotations, not main content. Reject purely decorative visuals with core information dependent on text lists\n   - Reject the inefficient pattern of \"large white space + centered single line of text\"\n4. **Narrative Flow & Chapters** — How should viewers move through the content? How is slide flow arranged?\n5. **Text Style & Density**\n   - Language: Explanatory text uses language explicitly requested by user, otherwise match user's conversation language\n   - Typography: Chinese and English titles preferably use serif fonts (Chinese uses Song font family)\n6. **Visual Style, Color & Mood**\n   - Visual language of encyclopedias and reference books: explanatory diagrams, cutaway illustrations, annotated structures\n   - Refined spatial composition and typographic precision of high-end journals\n   - Intentional asymmetry and layered information design of contemporary design publications\n   - Apply asymmetric grids, intentional breathing space, layered information organization, diagonal composition, dynamic typography as internalized design language\n   - **Color restriction:** Unless user explicitly specifies, do NOT use blue or purple as theme color or background color\n\n### Slides Workflow\n\n**Step 1: Design Strategy — Create Content Script**\n\nInformation architecture first: Structure content into hierarchical slides, each slide as an information unit defined by what data/facts/relationships it carries. Let content volume naturally determine slide count.\n\nOutput `content_script.md`:\n\n```\n# Slides Content Script\n\n## Slide 1: [Title]\n**Subtopic A**: [Label]\n[50-80 word narrative paragraph describing information content to be visualized]\n\n**Subtopic B**: [Label]\n[50-80 word narrative paragraph]\n\n## Slide 2: [Title]\n...\n```\n\nContent Script Specification:\n- Only describe \"what information needs to be presented\", not \"how to present it\"\n- Do NOT include \"Visual Description\" sections\n- Do NOT describe colors, backgrounds, decorative elements, atmosphere effects, mood, or layout details\n- Focus on pure information architecture\n- 2-3 focused subtopics per slide\n\n**Step 2: Sequential Image Generation**\n\nUse image generation tool to generate slides one by one:\n- First slide: Use gen_images (create from scratch)\n- Subsequent slides: Use edit_images, base_image_file points to previous slide\n\nFormat: Default 16:9 landscape ratio. Save each slide image locally.\n\n**Prompt Construction for Each Slide — Must include these 6 points:**\n\n1. **Visualization Type** — Prioritize diagram forms over text-dominated presentations: cutaway views, flowcharts, annotated structure diagrams, relationship diagrams, timeline overlays. Integrate multiple subtopics into unified visual structure. Avoid \"parallel cards/grid displays/multi-column layouts\" and text-heavy traditional typography.\n\n2. **Information Hierarchy** — Primary and secondary information distinguished through visual hierarchy (size, position, contrast). Not flat lists.\n\n3. **Composition Instructions** — Asymmetric layout, diagonal momentum, and other methods to break rigid symmetry.\n\n4. **Density Requirements** — Clear information hierarchy over quantity. Appropriate white space serves readability, but not empty and sparse.\n\n5. **Layout Independence** — Explicitly state this slide's visualization type is chosen based on its content, not copying previous slide. Re-evaluate what this specific content needs. But describe inherited elements in detail.\n\n6. **Style Consistency** — If user provided visual style or reference images, each prompt must describe that style's characteristics in detail.\n\n**Step 3: Compile Output**\n\nAfter generating all slide images:\n- Auto-compile into PDF (150 DPI, 95% quality, controlled file size)\n- Auto-compile into PPTX presentation\n\n### Slides Output\n- Files: `{topic}_slides.pdf` + `{topic}_slides.pptx`\n- Only return file paths, no preview in conversation\n- Do not output individual slide images, summary documents, content outlines, design descriptions, or usage instructions\n\n---\n\n## Format 4: Mind Map Generation\n\n### Mind Map Workflow\n\n**Step 1: Design Content Structure**\n\nDetermine node hierarchy and relationships:\n- Root node: Chapter theme\n- Level 1 nodes: Core concepts\n- Level 2 nodes: Detail points\n- No more than 4 levels\n- Each node text concise (no more than 10 characters)\n- Mark relationships between concepts (parallel/progressive/causal/contrast)\n\n**Step 2: Generate Image**\n\nUse gen_images to generate mind map image:\n- Format: 16:9 or square (based on content)\n- Style: Clear visual hierarchy, professional infographic style\n\n**Step 3: Output**\n\n- Mind map image: `{topic}_mindmap.png`\n- Attached Mermaid format text (optional, for users who need to edit)\n- Only return file path, no image preview in conversation\n\n---\n\n## Format 5: Audio Course Generation\n\n### Audio Workflow\n\n**Step 1: Write Dialogue Script**\n\nWrite teacher-student dialogue script:\n\n```\nOpening (about 1 minute)\n- Teacher greets, introduces today's topic\n- Student responds, expresses existing knowledge or questions\n- Teacher builds connection using user's interest field\n\nPart One: Concept Introduction (about 4 minutes)\n- Teacher asks questions from user's interest scenario\n- Student observes/answers\n- Teacher introduces core concept, defines in conversational manner\n- Student requests examples\n- Teacher explains in detail with personalized examples\n- Student restates in own words to confirm understanding\n\nPart Two: Deep Understanding (about 5 minutes)\n- Teacher explains important characteristics of concept\n- Student raises common confusion/misconception\n- Teacher clarifies misconception\n- Student poses hypothetical questions\n- Teacher answers and extends\n\nPart Three: Application Practice (about 3 minutes)\n- Teacher gives question\n- Student thinks and answers\n- Teacher provides feedback (affirmation or guidance)\n\nSummary (about 2 minutes)\n- Student attempts to summarize what was learned\n- Teacher supplements and affirms\n- Student expresses gains, connects to practical application\n- Exchange farewells\n```\n\nScript Requirements:\n- Dialogue natural, matches real teacher-student conversation rhythm\n- Avoid written expression\n- Include interjections (\"um\", \"well\", \"oh right\")\n- Allow student to \"interrupt\" with questions\n- All examples sourced from user's interest field\n- About 150-180 words per minute\n\n### Character Settings\n\n**Teacher Character:**\n- Professional yet approachable\n- Good at using metaphors to explain complex concepts\n- Patient in answering questions\n- Timely encouragement and affirmation\n\n**Student Character:**\n- Curious, actively asks questions\n- Represents target user's perspective\n- Makes common mistakes, raises typical confusions\n- Has own interest background (consistent with user settings)\n\n**Step 2: Generate Audio**\n\nUse audio generation tool to convert script to audio:\n- Teacher voice: Warm, professional, patient\n- Student voice: Curious, lively, sincere\n- Speed: Medium for concept explanation, natural rhythm for dialogue, slightly faster for summary\n\n**Step 3: Output**\n- File: `{topic}_audio.mp3`\n- Only return file path, no preview or playback in conversation\n- No auto-play\n- Do not output script files or production notes\n\n### Audio Quality Standards\n1. **Listening Experience** — Sounds like real conversation, not script reading; rhythm varies; key content emphasized\n2. **Learning Effect** — Concept explanation clear; student questions represent real confusion; practice section has testing effect\n3. **Personalization** — Examples 100% from user's interest field; student character gives user identification; language style matches grade\n4. **Audio Quality** — Clear sound; duration about 15 minutes; directly playable\n\n---\n\n## Critical Constraints\n\n1. **Content Fidelity** — All content must be based on original textbook/source material. No unverified information added.\n2. **Grade Adaptation** — Adjust content depth and expression based on grade level for ALL formats.\n3. **Output Rules** — Only return file paths. No inline display of images/PDFs/audio/video. No auto-play. No intermediate files.\n4. **Color Constraints (Notes)** — Maximum 2 colors per PDF. Default blue + pink.\n5. **Color Constraints (Slides)** — Do NOT use blue or purple as theme/background color unless user explicitly requests.\n6. **Image Quality** — Notes: 150 DPI, 90% compression. Slides: 150 DPI, 95% quality.\n7. **Mind Map Depth** — No more than 4 levels. Node text no more than 10 characters.\n8. **Quiz Minimalism** — No UI frameworks, no decorative elements, system default font only.\n\n## Common Mistakes to Avoid\n\n1. **Adding unverified information** — Stick to the source material only\n2. **Ignoring grade level** — Elementary content should not use university-level terminology\n3. **Previewing outputs in conversation** — Never display images, PDFs, or play audio inline\n4. **Dense annotations on notes** — Keep 3-8 annotations per page, not more\n5. **Decorative slides** — Visuals must convey information through structure, not just atmosphere\n6. **Text-heavy slides** — Diagrams should be primary carriers, not text lists with decorative backgrounds\n7. **Using blue/purple in slides** — Forbidden unless user explicitly requests\n8. **Flat quiz feedback** — \"The answer is X\" has no teaching value; always explain why\n9. **Robotic audio dialogue** — Must sound like natural conversation with interjections and interruptions\n10. **Outputting intermediate files** — Only deliver final output file paths\n\n## File & Output Conventions\n\n| Format | Filename Pattern | File Type |\n|--------|-----------------|-----------|\n| Notes | `{topic}_notes.pdf` | PDF |\n| Quiz | `{topic}_quiz.html` | HTML |\n| Slides | `{topic}_slides.pdf`, `{topic}_slides.pptx` | PDF, PPTX |\n| Mind Map | `{topic}_mindmap.png` | PNG |\n| Audio | `{topic}_audio.mp3` | MP3 |\n\nAll files use the topic name as prefix. Deliver all outputs together using `<deliver_assets>` format after all generation is complete.\n\nFile v11.0.0:_meta.json\n\n{\n  \"ownerId\": \"kn712egma3976f4b68hksznyrs85wsr5\",\n  \"slug\": \"eric-knowledge-digest\",\n  \"version\": \"11.0.0\",\n  \"publishedAt\": 1777802484635\n}\n\nFile v11.0.0:skill-card.md\n\n## Description:\n\nConverts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, and mind maps.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[ericn26-star](https://clawhub.ai/user/ericn26-star)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nExternal learners, educators, and developers can use this skill to convert PDFs, textbooks, or topic descriptions into grade-adapted study materials such as notes, quizzes, slides, mind maps, and audio lessons.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: Uploaded documents may contain private or sensitive material that the skill will parse and transform into study assets.\n\nMitigation: Use the skill only with documents intended for learning-material generation, and avoid unrelated private documents.\n\nRisk: Topic-only requests may use searched educational material before downstream study assets are generated.\n\nMitigation: Review and confirm compiled source material before proceeding with generated notes, quizzes, slides, mind maps, or audio.\n\n## Reference(s):\n\n- [ClawHub skill page](https://clawhub.ai/ericn26-star/skills/eric-knowledge-digest)\n\n## Skill Output:\n\n**Output Type(s):** [text, markdown, code, files, guidance]\n\n**Output Format:** [Markdown deliver_assets block containing local file paths for PDFs, PPTX, HTML, PNG, MP3, and optional Mermaid text]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [Final assets are returned as file paths only; previews, autoplay, and intermediate files are avoided.]\n\n## Skill Version(s):\n\n11.0.0 (source: ClawHub release metadata)\n\n## Ethical Considerations:\n\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.\n\nArchive v10.0.0: 2 files, 10472 bytes\n\nFiles: _meta.json (141b), SKILL.md (26234b)\n\nFile v10.0.0:SKILL.md\n\n---\nname: knowledge-digest\ndescription: \"Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, and mind maps. Trigger: learning materials, convert textbook, study notes, quiz generation, slides from PDF, mind map, audio course.\"\n---\n\n# KnowledgeDigest — Unified Learning Content Converter\n\n## Overview\n\nKnowledgeDigest converts textbooks, PDFs, or topic descriptions into personalized, multimodal learning experiences. It analyzes source content, then generates any combination of: handwritten-style notes (PDF), interactive quiz webpages (HTML), slides (PDF+PPTX), mind maps (image+Mermaid), and audio courses (MP3). All output is adapted to the learner's grade level and interests.\n\n## Workflow\n\n### Phase 1: Gather User Input\n\n1. Identify what the user has provided:\n   - Uploaded PDF/textbook file (optional)\n   - Topic/direction description\n   - Grade level (elementary / middle school / high school / university / professional)\n   - Expected output format(s)\n\n2. If no PDF/textbook uploaded and no source materials specified (only topic/direction provided):\n   - Ask user:\n     - **Option A:** \"I have materials, uploading now\"\n     - **Option B:** \"No materials, please search and generate courseware about [topic]\"\n   - If user selects B:\n     - Use search tools to collect authoritative materials on the topic\n     - Organize into structured content, generate a basic courseware PDF\n     - Send PDF to user for confirmation: \"This is the basic material I compiled for [topic], please confirm if usable?\"\n     - Continue after user confirmation\n\n3. **Default output formats** (if user does not specify): mindmap + slides (PDF only) + quiz\n\n### Phase 2: Content Analysis\n\nParse the PDF or structured content to extract:\n\n**Document Parsing:**\n- Identify chapter structure (chapters, sections, subsections)\n- Extract heading hierarchy and table of contents\n- Identify body text, images, tables, formulas, and other elements\n\n**Core Concept Extraction:**\n- Identify core concepts and key terms in each chapter\n- Extract definitions, theorems, formulas, and important content\n- Mark difficult points and key knowledge\n\n**Learning Objective Analysis:**\n- Infer learning objectives for each chapter\n- Identify prerequisite knowledge requirements\n- Analyze dependencies between knowledge points\n\n**Output structured analysis results in this format:**\n\n```json\n{\n  \"document_info\": {\n    \"title\": \"Document title\",\n    \"total_pages\": 100,\n    \"language\": \"zh/en\",\n    \"subject\": \"Subject area\"\n  },\n  \"chapters\": [\n    {\n      \"chapter_id\": \"1\",\n      \"title\": \"Chapter title\",\n      \"page_range\": [1, 20],\n      \"sections\": [\n        {\n          \"section_id\": \"1.1\",\n          \"title\": \"Section title\",\n          \"core_concepts\": [\"Concept 1\", \"Concept 2\"],\n          \"key_terms\": [\n            {\"term\": \"Term\", \"definition\": \"Definition\"}\n          ],\n          \"learning_objectives\": [\"Objective 1\", \"Objective 2\"],\n          \"difficulty\": \"easy/medium/hard\",\n          \"prerequisites\": [\"Prerequisite knowledge\"]\n        }\n      ]\n    }\n  ],\n  \"knowledge_graph\": {\n    \"nodes\": [\"Concept node list\"],\n    \"edges\": [{\"from\": \"Concept A\", \"to\": \"Concept B\", \"relation\": \"depends/contains/related\"}]\n  }\n}\n```\n\n**Parsing Rules:**\n1. Chapter Recognition — Identify hierarchy based on font size, bold, numbering, etc. Handle documents without clear chapter markers by logically segmenting.\n2. Concept Extraction — Identify bolded, highlighted, boxed important content. Extract proper nouns and term definitions. Identify formulas and theorems.\n3. Difficulty Assessment — Assess based on concept abstraction level, prerequisite knowledge, and content complexity.\n4. Quality Assurance — Ensure all chapters identified, verify knowledge point coverage completeness, check accuracy of concept definitions.\n\n### Phase 3: Generate Requested Formats\n\nBased on user-selected output formats, generate each in sequence. For each format, follow the corresponding section below.\n\n### Phase 4: Deliver Assets\n\nAfter all generation is complete:\n- Only return file paths, no previews allowed\n- No inline display of images/PDFs/audio/video in conversation\n- Audio/video files must not auto-play\n\nPresent to user using deliver_assets format:\n```\n<deliver_assets>\n<item>\n<path>file path</path>\n</item>\n</deliver_assets>\n```\n\n---\n\n## Supported Output Formats\n\n| Format | Output | Description |\n|--------|--------|-------------|\n| `notes` | `{topic}_notes.pdf` | Handwritten-style notes (annotated on original or generated from scratch) |\n| `quiz` | `{topic}_quiz.html` | Minimalist interactive HTML quiz with instant feedback |\n| `slides` | `{topic}_slides.pdf` + `{topic}_slides.pptx` | Visual slides |\n| `mindmap` | `{topic}_mindmap.png` + Mermaid text | Mind map image |\n| `audio` | `{topic}_audio.mp3` | Audio course in teacher-student dialogue format |\n| `all` | All of the above | Generate every format |\n\n---\n\n## Personalization: Grade Level Adaptation\n\nAll generated content must be adapted to the learner's grade level:\n\n| Grade | Language & Tone | Content Density | Visual Style |\n|-------|----------------|-----------------|--------------|\n| **Elementary** | Lively, simple Q&A, encouraging, story-style | Low density, more drawings, large font | Fun elements, bright colors, short text |\n| **Middle school** | Guided questioning, moderate challenges, youth-oriented | Moderate, image-text combination, clear labels | Image-text combination, moderate information |\n| **High school** | In-depth discussion, logical reasoning, appropriate academic tone | Higher density, logic diagrams | Professional feel, data visualization |\n| **University/Professional** | Seminar-style, critical thinking, professional terminology | High density, professional charts, complex structures | Academic style, comprehensive application |\n\n**Interest Adaptation** (applies to all formats):\n- Examples and metaphors use the user's interest field\n- Scenarios drawn from the user's familiar domain\n- Visual style and analogies match user interests\n\n---\n\n## Format 1: Notes Generation\n\n### Input Type Determination\n\n**Type A — Existing Paper/Courseware:**\n- PDF format academic papers, courseware/PPT exports, scanned textbook pages\n- Features: Fixed layout, page numbers, chapter numbering, formulas/charts\n- Action: Overlay handwritten notes on original pages\n\n**Type B — Non-existing Content:**\n- Plain text notes, knowledge point lists, oral transcripts, web content excerpts\n- Features: No fixed layout, needs reorganization\n- Action: Generate notes PDF from scratch\n\n### Type A Workflow: Adding Notes to Original Document\n\n**Step 1: Analyze Original Structure**\n\nAnalyze PDF content page by page:\n- Identify chapter titles and positions\n- Identify core concepts/terms\n- Identify formulas and their meanings\n- Identify problem/challenge statements\n- Identify solutions/methods\n- Identify key conclusions\n\n**Step 2: Plan Note Content**\n\nPlan handwritten annotations for each page (3-8 annotations per page, not too dense):\n\nAnnotation Types:\n1. **Chapter title translation/explanation** — e.g., original \"3.1 Preliminaries\" → annotate \"Background Knowledge\"\n2. **Key questions** — e.g., \"Key: How to reduce complexity?\"\n3. **Concept explanation** — e.g., annotate \"kernel trick\" next to formula\n4. **Problem marking** — e.g., \"Problem: memory overflow\"\n5. **Solutions** — e.g., \"Solution: forget gate\"\n6. **Formula notes** — e.g., \"recursive form\", \"write operation & read operation\"\n7. **Structure annotation** — e.g., use braces to mark formula groups, write \"→ O(N²) complexity\" beside\n\nAnnotation Planning Principles:\n- Positions avoid blocking key content\n- Utilize margins and paragraph gaps\n- Related content connected with lines or arrows\n\n**Step 3: Generate Annotated Images**\n\nConvert each PDF page to image, then use image generation tool to add handwritten-style annotations.\n\nHandwritten Annotation Style Requirements:\n- **Font:** Handwritten style, slightly tilted\n- **Color:** Unified colors throughout PDF, no more than 2\n  - Default: blue and pink (unless user specifies otherwise)\n  - All subsequent pages can only choose from these 2 colors\n  - Color assignment rules:\n    - Color 1 (blue/primary): Chapter titles, structure annotations, concept explanations, formula notes\n    - Color 2 (pink/accent): Key questions, problem marking, solutions\n- **Size:** Slightly larger than body text, eye-catching but not overwhelming\n- **Position:** Margins, paragraph gaps, blank space next to formulas\n\n**Step 4: Compile PDF**\n- Maintain original page order\n- Image quality: 150 DPI\n- Compression quality: 90%\n\n### Type B Workflow: Generating Notes from Scratch\n\n**Step 1: Organize Content Structure**\n- Main title → Chapters/modules → Core concepts → Key points/details → Examples/applications\n\n**Step 2: Design Note Layout**\n\nLayout Elements:\n- Title area: Large handwritten title\n- Body area: Handwritten-style bullet points\n- Diagram area: Concept maps, flowcharts, relationship diagrams (hand-drawn style)\n- Annotation area: Key markers, question marks, exclamation marks\n- Blank area: Space reserved for user's own notes\n\n**Step 3: Generate Note Page Images**\n\nEach page contains:\n- Page title (handwritten large text)\n- Core content (handwritten bullet points)\n- Diagrams (hand-drawn style concept maps/flowcharts)\n- Key annotations (boxes, arrows, underlines)\n- Notes (like \"Important!\", \"Common mistake\", \"Remember this\")\n\nStyle Requirements:\n- **Overall:** Looks like carefully made student notes, not printed document\n- **Font:** Handwritten, varying sizes (large for titles, medium for body, small for notes)\n- **Color:** Unified colors throughout PDF, no more than 2\n  - Default: blue and pink (unless user specifies otherwise)\n  - Color assignment: Blue (titles, framework, notes), Pink (key points)\n- **Layout:** Organized but not rigid, slight tilting and variation allowed\n- **Elements:** Arrows, underlines, boxes, cloud frames, asterisks — use only when necessary\n\n**Step 4: Compile PDF**\n- Arrange in logical content order\n- Image quality: 150 DPI, compression quality: 90%\n\n### Notes Output\n- File: `{topic}_notes.pdf`\n- Only return file path, no preview in conversation\n- Do not output intermediate image files or content scripts\n\n### Notes Quality Standards\n1. **Content Accuracy** — Annotations based on original text; translation/explanation accurate; no added information\n2. **Annotation Value** — Annotations help understanding, not simple repetition; key points highlight important concepts; problems and solutions correspond clearly\n3. **Visual Effect** — Handwritten style natural, not machine-printed; color coordination harmonious; annotation positions reasonable\n4. **Usability** — PDF printable; suitable for screen reading; reasonable file size\n\n---\n\n## Format 2: Quiz Generation\n\n### Question Design\n\nAt least 5 questions per section. Distribution:\n- Multiple choice (multiple_choice): 2-3 questions\n- True/false (true_false): 1-2 questions\n- Fill in the blank (fill_blank): 1-2 questions\n\nDifficulty distribution:\n- 40% Easy (memory, comprehension)\n- 40% Medium (application)\n- 20% Hard (analysis, synthesis)\n\nEach question must include:\n- Question content (using personalized scenario)\n- Correct answer\n- Answer explanation (has teaching value, not just \"the answer is X\")\n- Related core concept\n\n### HTML Generation\n\nGenerate a single HTML file containing all questions and interaction logic.\n\n**Design Principle: Minimalist**\n\nVisual Style:\n- Pure white background\n- Black text\n- No decorative elements, no icons, no gradients, no shadows\n- No borders or only 1px gray thin lines\n- Font: System default font\n- Minimal CSS, no UI frameworks\n\nInteraction Design:\n- Click option to select, selected state distinguished by slight background color\n- Show correct/incorrect and explanation immediately after submit\n- Correct: Green text \"Correct\"\n- Incorrect: Red text \"Incorrect\" + correct answer + explanation\n- Show total score at end\n\n**HTML Structure Template:**\n\n```html\n<!DOCTYPE html>\n<html lang=\"en\">\n<head>\n  <meta charset=\"UTF-8\">\n  <meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">\n  <title>Chapter Quiz</title>\n  <style>\n    body {\n      font-family: system-ui, sans-serif;\n      max-width: 600px;\n      margin: 40px auto;\n      padding: 20px;\n      line-height: 1.6;\n    }\n    h1 { font-size: 1.5em; font-weight: normal; }\n    .question { margin: 30px 0; }\n    .question-text { margin-bottom: 15px; }\n    .option {\n      display: block;\n      padding: 10px;\n      margin: 5px 0;\n      cursor: pointer;\n    }\n    .option:hover { background: #f5f5f5; }\n    .option.selected { background: #e8e8e8; }\n    .feedback { margin-top: 10px; font-size: 0.9em; }\n    .correct { color: #2e7d32; }\n    .incorrect { color: #c62828; }\n    .explanation { color: #666; margin-top: 5px; }\n    button {\n      padding: 10px 20px;\n      background: #333;\n      color: white;\n      border: none;\n      cursor: pointer;\n      margin-top: 20px;\n    }\n    .score { font-size: 1.2em; margin-top: 30px; }\n  </style>\n</head>\n<body>\n  <h1>Chapter Title - Quiz</h1>\n\n  <div class=\"question\" data-answer=\"A\">\n    <div class=\"question-text\">1. Question content</div>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"A\"> A. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"B\"> B. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"C\"> C. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"D\"> D. Option</label>\n    <div class=\"feedback\"></div>\n  </div>\n\n  <!-- More questions... -->\n\n  <button onclick=\"submit()\">Submit</button>\n  <div class=\"score\"></div>\n\n  <script>\n    const explanations = {\n      q1: \"Explanation content...\",\n      // ...\n    };\n\n    function submit() {\n      let correct = 0;\n      document.querySelectorAll('.question').forEach((q, i) => {\n        const answer = q.dataset.answer;\n        const selected = q.querySelector('input:checked');\n        const feedback = q.querySelector('.feedback');\n        const qName = 'q' + (i + 1);\n\n        if (selected && selected.value === answer) {\n          feedback.innerHTML = '<span class=\"correct\">Correct</span>';\n          correct++;\n        } else {\n          feedback.innerHTML = '<span class=\"incorrect\">Incorrect</span> Correct answer: ' + answer +\n            '<div class=\"explanation\">' + explanations[qName] + '</div>';\n        }\n      });\n\n      document.querySelector('.score').textContent =\n        'Score: ' + correct + '/' + document.querySelectorAll('.question').length;\n    }\n  </script>\n</body>\n</html>\n```\n\n### Quiz Output\n- File: `{topic}_quiz.html`\n- Only return file path, no preview in conversation\n- Do not output JSON data, CSS files, or JS files separately\n\n### Quiz Quality Standards\n1. **Content Accuracy** — All knowledge points based on original textbook; answers and explanations correct; question wording clear and unambiguous\n2. **Personalization** — Question scenarios match user interests; difficulty matches grade level; language style suits target audience\n3. **Interaction Experience** — Click response instant; feedback clear; explanations have teaching value\n4. **Visual Minimalism** — No decorative elements; no framework dependencies; file size minimized\n\n---\n\n## Format 3: Slides Generation\n\n### Design Considerations\n\nTreat these as a flexible menu, not a mandatory checklist:\n\n1. **Topic, Purpose & Audience** — What is this about? Who needs to understand it? Where will it be presented?\n2. **Content Foundation & Sources** — What materials or data need to be presented?\n3. **Visual Approach (CRITICAL)**\n   - Default to explanatory visuals: cutaway views, annotated structure diagrams, exploded views, schematic illustrations\n   - Visual elements are primary information carriers, not decorative backgrounds for text lists\n   - Default information density matches professional infographics and technical illustrations\n   - **CRITICAL:** Diagrams must convey information through structure, not just provide atmosphere. Text should be labels/annotations, not main content. Reject purely decorative visuals with core information dependent on text lists\n   - Reject the inefficient pattern of \"large white space + centered single line of text\"\n4. **Narrative Flow & Chapters** — How should viewers move through the content? How is slide flow arranged?\n5. **Text Style & Density**\n   - Language: Explanatory text uses language explicitly requested by user, otherwise match user's conversation language\n   - Typography: Chinese and English titles preferably use serif fonts (Chinese uses Song font family)\n6. **Visual Style, Color & Mood**\n   - Visual language of encyclopedias and reference books: explanatory diagrams, cutaway illustrations, annotated structures\n   - Refined spatial composition and typographic precision of high-end journals\n   - Intentional asymmetry and layered information design of contemporary design publications\n   - Apply asymmetric grids, intentional breathing space, layered information organization, diagonal composition, dynamic typography as internalized design language\n   - **Color restriction:** Unless user explicitly specifies, do NOT use blue or purple as theme color or background color\n\n### Slides Workflow\n\n**Step 1: Design Strategy — Create Content Script**\n\nInformation architecture first: Structure content into hierarchical slides, each slide as an information unit defined by what data/facts/relationships it carries. Let content volume naturally determine slide count.\n\nOutput `content_script.md`:\n\n```\n# Slides Content Script\n\n## Slide 1: [Title]\n**Subtopic A**: [Label]\n[50-80 word narrative paragraph describing information content to be visualized]\n\n**Subtopic B**: [Label]\n[50-80 word narrative paragraph]\n\n## Slide 2: [Title]\n...\n```\n\nContent Script Specification:\n- Only describe \"what information needs to be presented\", not \"how to present it\"\n- Do NOT include \"Visual Description\" sections\n- Do NOT describe colors, backgrounds, decorative elements, atmosphere effects, mood, or layout details\n- Focus on pure information architecture\n- 2-3 focused subtopics per slide\n\n**Step 2: Sequential Image Generation**\n\nUse image generation tool to generate slides one by one:\n- First slide: Use gen_images (create from scratch)\n- Subsequent slides: Use edit_images, base_image_file points to previous slide\n\nFormat: Default 16:9 landscape ratio. Save each slide image locally.\n\n**Prompt Construction for Each Slide — Must include these 6 points:**\n\n1. **Visualization Type** — Prioritize diagram forms over text-dominated presentations: cutaway views, flowcharts, annotated structure diagrams, relationship diagrams, timeline overlays. Integrate multiple subtopics into unified visual structure. Avoid \"parallel cards/grid displays/multi-column layouts\" and text-heavy traditional typography.\n\n2. **Information Hierarchy** — Primary and secondary information distinguished through visual hierarchy (size, position, contrast). Not flat lists.\n\n3. **Composition Instructions** — Asymmetric layout, diagonal momentum, and other methods to break rigid symmetry.\n\n4. **Density Requirements** — Clear information hierarchy over quantity. Appropriate white space serves readability, but not empty and sparse.\n\n5. **Layout Independence** — Explicitly state this slide's visualization type is chosen based on its content, not copying previous slide. Re-evaluate what this specific content needs. But describe inherited elements in detail.\n\n6. **Style Consistency** — If user provided visual style or reference images, each prompt must describe that style's characteristics in detail.\n\n**Step 3: Compile Output**\n\nAfter generating all slide images:\n- Auto-compile into PDF (150 DPI, 95% quality, controlled file size)\n- Auto-compile into PPTX presentation\n\n### Slides Output\n- Files: `{topic}_slides.pdf` + `{topic}_slides.pptx`\n- Only return file paths, no preview in conversation\n- Do not output individual slide images, summary documents, content outlines, design descriptions, or usage instructions\n\n---\n\n## Format 4: Mind Map Generation\n\n### Mind Map Workflow\n\n**Step 1: Design Content Structure**\n\nDetermine node hierarchy and relationships:\n- Root node: Chapter theme\n- Level 1 nodes: Core concepts\n- Level 2 nodes: Detail points\n- No more than 4 levels\n- Each node text concise (no more than 10 characters)\n- Mark relationships between concepts (parallel/progressive/causal/contrast)\n\n**Step 2: Generate Image**\n\nUse gen_images to generate mind map image:\n- Format: 16:9 or square (based on content)\n- Style: Clear visual hierarchy, professional infographic style\n\n**Step 3: Output**\n\n- Mind map image: `{topic}_mindmap.png`\n- Attached Mermaid format text (optional, for users who need to edit)\n- Only return file path, no image preview in conversation\n\n---\n\n## Format 5: Audio Course Generation\n\n### Audio Workflow\n\n**Step 1: Write Dialogue Script**\n\nWrite teacher-student dialogue script:\n\n```\nOpening (about 1 minute)\n- Teacher greets, introduces today's topic\n- Student responds, expresses existing knowledge or questions\n- Teacher builds connection using user's interest field\n\nPart One: Concept Introduction (about 4 minutes)\n- Teacher asks questions from user's interest scenario\n- Student observes/answers\n- Teacher introduces core concept, defines in conversational manner\n- Student requests examples\n- Teacher explains in detail with personalized examples\n- Student restates in own words to confirm understanding\n\nPart Two: Deep Understanding (about 5 minutes)\n- Teacher explains important characteristics of concept\n- Student raises common confusion/misconception\n- Teacher clarifies misconception\n- Student poses hypothetical questions\n- Teacher answers and extends\n\nPart Three: Application Practice (about 3 minutes)\n- Teacher gives question\n- Student thinks and answers\n- Teacher provides feedback (affirmation or guidance)\n\nSummary (about 2 minutes)\n- Student attempts to summarize what was learned\n- Teacher supplements and affirms\n- Student expresses gains, connects to practical application\n- Exchange farewells\n```\n\nScript Requirements:\n- Dialogue natural, matches real teacher-student conversation rhythm\n- Avoid written expression\n- Include interjections (\"um\", \"well\", \"oh right\")\n- Allow student to \"interrupt\" with questions\n- All examples sourced from user's interest field\n- About 150-180 words per minute\n\n### Character Settings\n\n**Teacher Character:**\n- Professional yet approachable\n- Good at using metaphors to explain complex concepts\n- Patient in answering questions\n- Timely encouragement and affirmation\n\n**Student Character:**\n- Curious, actively asks questions\n- Represents target user's perspective\n- Makes common mistakes, raises typical confusions\n- Has own interest background (consistent with user settings)\n\n**Step 2: Generate Audio**\n\nUse audio generation tool to convert script to audio:\n- Teacher voice: Warm, professional, patient\n- Student voice: Curious, lively, sincere\n- Speed: Medium for concept explanation, natural rhythm for dialogue, slightly faster for summary\n\n**Step 3: Output**\n- File: `{topic}_audio.mp3`\n- Only return file path, no preview or playback in conversation\n- No auto-play\n- Do not output script files or production notes\n\n### Audio Quality Standards\n1. **Listening Experience** — Sounds like real conversation, not script reading; rhythm varies; key content emphasized\n2. **Learning Effect** — Concept explanation clear; student questions represent real confusion; practice section has testing effect\n3. **Personalization** — Examples 100% from user's interest field; student character gives user identification; language style matches grade\n4. **Audio Quality** — Clear sound; duration about 15 minutes; directly playable\n\n---\n\n## Critical Constraints\n\n1. **Content Fidelity** — All content must be based on original textbook/source material. No unverified information added.\n2. **Grade Adaptation** — Adjust content depth and expression based on grade level for ALL formats.\n3. **Output Rules** — Only return file paths. No inline display of images/PDFs/audio/video. No auto-play. No intermediate files.\n4. **Color Constraints (Notes)** — Maximum 2 colors per PDF. Default blue + pink.\n5. **Color Constraints (Slides)** — Do NOT use blue or purple as theme/background color unless user explicitly requests.\n6. **Image Quality** — Notes: 150 DPI, 90% compression. Slides: 150 DPI, 95% quality.\n7. **Mind Map Depth** — No more than 4 levels. Node text no more than 10 characters.\n8. **Quiz Minimalism** — No UI frameworks, no decorative elements, system default font only.\n\n## Common Mistakes to Avoid\n\n1. **Adding unverified information** — Stick to the source material only\n2. **Ignoring grade level** — Elementary content should not use university-level terminology\n3. **Previewing outputs in conversation** — Never display images, PDFs, or play audio inline\n4. **Dense annotations on notes** — Keep 3-8 annotations per page, not more\n5. **Decorative slides** — Visuals must convey information through structure, not just atmosphere\n6. **Text-heavy slides** — Diagrams should be primary carriers, not text lists with decorative backgrounds\n7. **Using blue/purple in slides** — Forbidden unless user explicitly requests\n8. **Flat quiz feedback** — \"The answer is X\" has no teaching value; always explain why\n9. **Robotic audio dialogue** — Must sound like natural conversation with interjections and interruptions\n10. **Outputting intermediate files** — Only deliver final output file paths\n\n## File & Output Conventions\n\n| Format | Filename Pattern | File Type |\n|--------|-----------------|-----------|\n| Notes | `{topic}_notes.pdf` | PDF |\n| Quiz | `{topic}_quiz.html` | HTML |\n| Slides | `{topic}_slides.pdf`, `{topic}_slides.pptx` | PDF, PPTX |\n| Mind Map | `{topic}_mindmap.png` | PNG |\n| Audio | `{topic}_audio.mp3` | MP3 |\n\nAll files use the topic name as prefix. Deliver all outputs together using `<deliver_assets>` format after all generation is complete.\n\nFile v10.0.0:_meta.json\n\n{\n  \"ownerId\": \"kn712egma3976f4b68hksznyrs85wsr5\",\n  \"slug\": \"eric-knowledge-digest\",\n  \"version\": \"10.0.0\",\n  \"publishedAt\": 1777784555298\n}\n\nArchive v3.0.1: 2 files, 10472 bytes\n\nFiles: _meta.json (140b), SKILL.md (26234b)\n\nFile v3.0.1:SKILL.md\n\n---\nname: knowledge-digest\ndescription: \"Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, and mind maps. Trigger: learning materials, convert textbook, study notes, quiz generation, slides from PDF, mind map, audio course.\"\n---\n\n# KnowledgeDigest — Unified Learning Content Converter\n\n## Overview\n\nKnowledgeDigest converts textbooks, PDFs, or topic descriptions into personalized, multimodal learning experiences. It analyzes source content, then generates any combination of: handwritten-style notes (PDF), interactive quiz webpages (HTML), slides (PDF+PPTX), mind maps (image+Mermaid), and audio courses (MP3). All output is adapted to the learner's grade level and interests.\n\n## Workflow\n\n### Phase 1: Gather User Input\n\n1. Identify what the user has provided:\n   - Uploaded PDF/textbook file (optional)\n   - Topic/direction description\n   - Grade level (elementary / middle school / high school / university / professional)\n   - Expected output format(s)\n\n2. If no PDF/textbook uploaded and no source materials specified (only topic/direction provided):\n   - Ask user:\n     - **Option A:** \"I have materials, uploading now\"\n     - **Option B:** \"No materials, please search and generate courseware about [topic]\"\n   - If user selects B:\n     - Use search tools to collect authoritative materials on the topic\n     - Organize into structured content, generate a basic courseware PDF\n     - Send PDF to user for confirmation: \"This is the basic material I compiled for [topic], please confirm if usable?\"\n     - Continue after user confirmation\n\n3. **Default output formats** (if user does not specify): mindmap + slides (PDF only) + quiz\n\n### Phase 2: Content Analysis\n\nParse the PDF or structured content to extract:\n\n**Document Parsing:**\n- Identify chapter structure (chapters, sections, subsections)\n- Extract heading hierarchy and table of contents\n- Identify body text, images, tables, formulas, and other elements\n\n**Core Concept Extraction:**\n- Identify core concepts and key terms in each chapter\n- Extract definitions, theorems, formulas, and important content\n- Mark difficult points and key knowledge\n\n**Learning Objective Analysis:**\n- Infer learning objectives for each chapter\n- Identify prerequisite knowledge requirements\n- Analyze dependencies between knowledge points\n\n**Output structured analysis results in this format:**\n\n```json\n{\n  \"document_info\": {\n    \"title\": \"Document title\",\n    \"total_pages\": 100,\n    \"language\": \"zh/en\",\n    \"subject\": \"Subject area\"\n  },\n  \"chapters\": [\n    {\n      \"chapter_id\": \"1\",\n      \"title\": \"Chapter title\",\n      \"page_range\": [1, 20],\n      \"sections\": [\n        {\n          \"section_id\": \"1.1\",\n          \"title\": \"Section title\",\n          \"core_concepts\": [\"Concept 1\", \"Concept 2\"],\n          \"key_terms\": [\n            {\"term\": \"Term\", \"definition\": \"Definition\"}\n          ],\n          \"learning_objectives\": [\"Objective 1\", \"Objective 2\"],\n          \"difficulty\": \"easy/medium/hard\",\n          \"prerequisites\": [\"Prerequisite knowledge\"]\n        }\n      ]\n    }\n  ],\n  \"knowledge_graph\": {\n    \"nodes\": [\"Concept node list\"],\n    \"edges\": [{\"from\": \"Concept A\", \"to\": \"Concept B\", \"relation\": \"depends/contains/related\"}]\n  }\n}\n```\n\n**Parsing Rules:**\n1. Chapter Recognition — Identify hierarchy based on font size, bold, numbering, etc. Handle documents without clear chapter markers by logically segmenting.\n2. Concept Extraction — Identify bolded, highlighted, boxed important content. Extract proper nouns and term definitions. Identify formulas and theorems.\n3. Difficulty Assessment — Assess based on concept abstraction level, prerequisite knowledge, and content complexity.\n4. Quality Assurance — Ensure all chapters identified, verify knowledge point coverage completeness, check accuracy of concept definitions.\n\n### Phase 3: Generate Requested Formats\n\nBased on user-selected output formats, generate each in sequence. For each format, follow the corresponding section below.\n\n### Phase 4: Deliver Assets\n\nAfter all generation is complete:\n- Only return file paths, no previews allowed\n- No inline display of images/PDFs/audio/video in conversation\n- Audio/video files must not auto-play\n\nPresent to user using deliver_assets format:\n```\n<deliver_assets>\n<item>\n<path>file path</path>\n</item>\n</deliver_assets>\n```\n\n---\n\n## Supported Output Formats\n\n| Format | Output | Description |\n|--------|--------|-------------|\n| `notes` | `{topic}_notes.pdf` | Handwritten-style notes (annotated on original or generated from scratch) |\n| `quiz` | `{topic}_quiz.html` | Minimalist interactive HTML quiz with instant feedback |\n| `slides` | `{topic}_slides.pdf` + `{topic}_slides.pptx` | Visual slides |\n| `mindmap` | `{topic}_mindmap.png` + Mermaid text | Mind map image |\n| `audio` | `{topic}_audio.mp3` | Audio course in teacher-student dialogue format |\n| `all` | All of the above | Generate every format |\n\n---\n\n## Personalization: Grade Level Adaptation\n\nAll generated content must be adapted to the learner's grade level:\n\n| Grade | Language & Tone | Content Density | Visual Style |\n|-------|----------------|-----------------|--------------|\n| **Elementary** | Lively, simple Q&A, encouraging, story-style | Low density, more drawings, large font | Fun elements, bright colors, short text |\n| **Middle school** | Guided questioning, moderate challenges, youth-oriented | Moderate, image-text combination, clear labels | Image-text combination, moderate information |\n| **High school** | In-depth discussion, logical reasoning, appropriate academic tone | Higher density, logic diagrams | Professional feel, data visualization |\n| **University/Professional** | Seminar-style, critical thinking, professional terminology | High density, professional charts, complex structures | Academic style, comprehensive application |\n\n**Interest Adaptation** (applies to all formats):\n- Examples and metaphors use the user's interest field\n- Scenarios drawn from the user's familiar domain\n- Visual style and analogies match user interests\n\n---\n\n## Format 1: Notes Generation\n\n### Input Type Determination\n\n**Type A — Existing Paper/Courseware:**\n- PDF format academic papers, courseware/PPT exports, scanned textbook pages\n- Features: Fixed layout, page numbers, chapter numbering, formulas/charts\n- Action: Overlay handwritten notes on original pages\n\n**Type B — Non-existing Content:**\n- Plain text notes, knowledge point lists, oral transcripts, web content excerpts\n- Features: No fixed layout, needs reorganization\n- Action: Generate notes PDF from scratch\n\n### Type A Workflow: Adding Notes to Original Document\n\n**Step 1: Analyze Original Structure**\n\nAnalyze PDF content page by page:\n- Identify\n\nArchive v3.0.0: 2 files, 10473 bytes\n\nFiles: _meta.json (140b), SKILL.md (26234b)\n\nArchive v2.0.0: 2 files, 10472 bytes\n\nFiles: _meta.json (140b), SKILL.md (26234b)\n\nArchive v1.0.2: 2 files, 10472 bytes\n\nFiles: _meta.json (140b), SKILL.md (26234b)\n\nArchive v1.0.1: 2 files, 10472 bytes\n\nFiles: _meta.json (140b), SKILL.md (26234b)","readmeExcerpt":"Skill: Knowledge Digest Owner: ericn26-star Summary: Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, an... Tags: latest:11.0.0 Version history: v1.2.0 | 2026-05-05T10:17:41.510Z | auto Version 1.2.0 — No changes detected in this release. - No file changes or updates were made compared to the previous version. v1","codeSnippets":[],"executableExamples":[{"language":"json","snippet":"{\n  \"document_info\": {\n    \"title\": \"Document title\",\n    \"total_pages\": 100,\n    \"language\": \"zh/en\",\n    \"subject\": \"Subject area\"\n  },\n  \"chapters\": [\n    {\n      \"chapter_id\": \"1\",\n      \"title\": \"Chapter title\",\n      \"page_range\": [1, 20],\n      \"sections\": [\n        {\n          \"section_id\": \"1.1\",\n          \"title\": \"Section title\",\n          \"core_concepts\": [\"Concept 1\", \"Concept 2\"],\n          \"key_terms\": [\n            {\"term\": \"Term\", \"definition\": \"Definition\"}\n          ],\n          \"learning_objectives\": [\"Objective 1\", \"Objective 2\"],\n          \"difficulty\": \"easy/medium/hard\",\n          \"prerequisites\": [\"Prerequisite knowledge\"]\n        }\n      ]\n    }\n  ],\n  \"knowledge_graph\": {\n    \"nodes\": [\"Concept node list\"],\n    \"edges\": [{\"from\": \"Concept A\", \"to\": \"Concept B\", \"relation\": \"depends/contains/related\"}]\n  }\n}"},{"language":"text","snippet":"<deliver_assets>\n<item>\n<path>file path</path>\n</item>\n</deliver_assets>"},{"language":"html","snippet":"<!DOCTYPE html>\n<html lang=\"en\">\n<head>\n  <meta charset=\"UTF-8\">\n  <meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">\n  <title>Chapter Quiz</title>\n  <style>\n    body {\n      font-family: system-ui, sans-serif;\n      max-width: 600px;\n      margin: 40px auto;\n      padding: 20px;\n      line-height: 1.6;\n    }\n    h1 { font-size: 1.5em; font-weight: normal; }\n    .question { margin: 30px 0; }\n    .question-text { margin-bottom: 15px; }\n    .option {\n      display: block;\n      padding: 10px;\n      margin: 5px 0;\n      cursor: pointer;\n    }\n    .option:hover { background: #f5f5f5; }\n    .option.selected { background: #e8e8e8; }\n    .feedback { margin-top: 10px; font-size: 0.9em; }\n    .correct { color: #2e7d32; }\n    .incorrect { color: #c62828; }\n    .explanation { color: #666; margin-top: 5px; }\n    button {\n      padding: 10px 20px;\n      background: #333;\n      color: white;\n      border: none;\n      cursor: pointer;\n      margin-top: 20px;\n    }\n    .score { font-size: 1.2em; margin-top: 30px; }\n  </style>\n</head>\n<body>\n  <h1>Chapter Title - Quiz</h1>\n\n  <div class=\"question\" data-answer=\"A\">\n    <div class=\"question-text\">1. Question content</div>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"A\"> A. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"B\"> B. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"C\"> C. Option</label>\n    <label class=\"option\"><input type=\"radio\" name=\"q1\" value=\"D\"> D. Option</label>\n    <div class=\"feedback\"></div>\n  </div>\n\n  <!-- More questions... -->\n\n  <button onclick=\"submit()\">Submit</button>\n  <div class=\"score\"></div>\n\n  <script>\n    const explanations = {\n      q1: \"Explanation content...\",\n      // ...\n    };\n\n    function submit() {\n      let correct = 0;\n      document.querySelectorAll('.question').forEach((q, i) => {\n        const answer = q.dataset.answer;\n        const selected = q.querySelector('input:checked');\n        const "},{"language":"text","snippet":"# Slides Content Script\n\n## Slide 1: [Title]\n**Subtopic A**: [Label]\n[50-80 word narrative paragraph describing information content to be visualized]\n\n**Subtopic B**: [Label]\n[50-80 word narrative paragraph]\n\n## Slide 2: [Title]\n..."},{"language":"text","snippet":"Opening (about 1 minute)\n- Teacher greets, introduces today's topic\n- Student responds, expresses existing knowledge or questions\n- Teacher builds connection using user's interest field\n\nPart One: Concept Introduction (about 4 minutes)\n- Teacher asks questions from user's interest scenario\n- Student observes/answers\n- Teacher introduces core concept, defines in conversational manner\n- Student requests examples\n- Teacher explains in detail with personalized examples\n- Student restates in own words to confirm understanding\n\nPart Two: Deep Understanding (about 5 minutes)\n- Teacher explains important characteristics of concept\n- Student raises common confusion/misconception\n- Teacher clarifies misconception\n- Student poses hypothetical questions\n- Teacher answers and extends\n\nPart Three: Application Practice (about 3 minutes)\n- Teacher gives question\n- Student thinks and answers\n- Teacher provides feedback (affirmation or guidance)\n\nSummary (about 2 minutes)\n- Student attempts to summarize what was learned\n- Teacher supplements and affirms\n- Student expresses gains, connects to practical application\n- Exchange farewells"},{"language":"json","snippet":"{\n  \"document_info\": {\n    \"title\": \"Document title\",\n    \"total_pages\": 100,\n    \"language\": \"zh/en\",\n    \"subject\": \"Subject area\"\n  },\n  \"chapters\": [\n    {\n      \"chapter_id\": \"1\",\n      \"title\": \"Chapter title\",\n      \"page_range\": [1, 20],\n      \"sections\": [\n        {\n          \"section_id\": \"1.1\",\n          \"title\": \"Section title\",\n          \"core_concepts\": [\"Concept 1\", \"Concept 2\"],\n          \"key_terms\": [\n            {\"term\": \"Term\", \"definition\": \"Definition\"}\n          ],\n          \"learning_objectives\": [\"Objective 1\", \"Objective 2\"],\n          \"difficulty\": \"easy/medium/hard\",\n          \"prerequisites\": [\"Prerequisite knowledge\"]\n        }\n      ]\n    }\n  ],\n  \"knowledge_graph\": {\n    \"nodes\": [\"Concept node list\"],\n    \"edges\": [{\"from\": \"Concept A\", \"to\": \"Concept B\", \"relation\": \"depends/contains/related\"}]\n  }\n}"}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[{"path":"SKILL.md","content":"---\nname: knowledge-digest\ndescription: \"Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, and mind maps. Trigger: learning materials, convert textbook, study notes, quiz generation, slides from PDF, mind map, audio course.\"\n---\n\n# KnowledgeDigest — Unified Learning Content Converter\n\n## Overview\n\nKnowledgeDigest converts textbooks, PDFs, or topic descriptions into personalized, multimodal learning experiences. It analyzes source content, then generates any combination of: handwritten-style notes (PDF), interactive quiz webpages (HTML), slides (PDF+PPTX), mind maps (image+Mermaid), and audio courses (MP3). All output is adapted to the learner's grade level and interests.\n\n## Workflow\n\n### Phase 1: Gather User Input\n\n1. Identify what the user has provided:\n   - Uploaded PDF/textbook file (optional)\n   - Topic/direction description\n   - Grade level (elementary / middle school / high school / university / professional)\n   - Expected output format(s)\n\n2. If no PDF/textbook uploaded and no source materials specified (only topic/direction provided):\n   - Ask user:\n     - **Option A:** \"I have materials, uploading now\"\n     - **Option B:** \"No materials, please search and generate courseware about [topic]\"\n   - If user selects B:\n     - Use search tools to collect authoritative materials on the topic\n     - Organize into structured content, generate a basic courseware PDF\n     - Send PDF to user for confirmation: \"This is the basic material I compiled for [topic], please confirm if usable?\"\n     - Continue after user confirmation\n\n3. **Default output formats** (if user does not specify): mindmap + slides (PDF only) + quiz\n\n### Phase 2: Content Analysis\n\nParse the PDF or structured content to extract:\n\n**Document Parsing:**\n- Identify chapter structure (chapters, sections, subsections)\n- Extract heading hierarchy and table of contents\n- Identify body text, images, tables, formulas, and other elements\n\n**Core Concept Extraction:**\n- Identify core concepts and key terms in each chapter\n- Extract definitions, theorems, formulas, and important content\n- Mark difficult points and key knowledge\n\n**Learning Objective Analysis:**\n- Infer learning objectives for each chapter\n- Identify prerequisite knowledge requirements\n- Analyze dependencies between knowledge points\n\n**Output structured analysis results in this format:**\n\n```json\n{\n  \"document_info\": {\n    \"title\": \"Document title\",\n    \"total_pages\": 100,\n    \"language\": \"zh/en\",\n    \"subject\": \"Subject area\"\n  },\n  \"chapters\": [\n    {\n      \"chapter_id\": \"1\",\n      \"title\": \"Chapter title\",\n      \"page_range\": [1, 20],\n      \"sections\": [\n        {\n          \"section_id\": \"1.1\",\n          \"title\": \"Section title\",\n          \"core_concepts\": [\"Concept 1\", \"Concept 2\"],\n          \"key_terms\": [\n            {\"term\": \"Term\", \"definition\": \"Definition\"}\n          ],\n          \"learning_objectives\": [\"Objective "},{"path":"_meta.json","content":"{\n  \"ownerId\": \"kn712egma3976f4b68hksznyrs85wsr5\",\n  \"slug\": \"eric-knowledge-digest\",\n  \"version\": \"1.2.0\",\n  \"publishedAt\": 1777976261510\n}"}],"languages":[],"docsSourceLabel":"CLAWHUB","editorialOverview":"Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, an... Skill: Knowledge Digest Owner: ericn26-star Summary: Converts textbooks or PDFs into personalized, multimodal interactive learning materials including handwritten notes, quiz webpages, slides, audio courses, an... Tags: latest:11.0.0 Version history: v1.2.0 | 2026-05-05T10:17:41.510Z | auto Version 1.2.0 — No changes detected in this release. - No file changes or updates were made compared to the previous version. v1","editorialQuality":{"score":100,"threshold":65,"status":"ready","wordCount":1013,"uniquenessScore":51,"reasons":[]}},"media":{"evidence":{"source":"no-media","verified":false,"confidence":"low","updatedAt":"2026-10-11T23:43:24.843Z","emptyReason":"No screenshots, media assets, or demo links are available."},"primaryImageUrl":null,"mediaAssetCount":0,"assets":[],"demoUrl":null},"ownerResources":{"evidence":{"source":"unclaimed","verified":false,"confidence":"low","updatedAt":"2026-10-11T23:43:24.843Z","emptyReason":"This page has not been claimed by the agent owner."},"hasCustomPage":false,"customPageUpdatedAt":null,"customLinks":[],"structuredLinks":{"docsUrl":null,"demoUrl":null,"supportUrl":null,"pricingUrl":null,"statusUrl":null},"customPage":null},"relatedAgents":{"evidence":{"source":"protocol-neighbors","verified":false,"confidence":"medium","updatedAt":"2026-10-13T00:33:30.275Z","emptyReason":null},"items":[{"id":"8ebccd8e-3863-4187-8355-c3f14e1f9edf","entityType":"agent","canonicalPath":"/agent/iofficeai-aionui","slug":"iofficeai-aionui","name":"AionUi","description":"Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!","url":"https://github.com/iOfficeAI/AionUi","homepage":"https://www.aionui.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-10-09T19:11:12.944Z","createdAt":"2026-02-25T03:38:16.584Z","downloads":null},{"id":"b917f68a-ebff-438e-84f8-3f4b2494c0bc","entityType":"agent","canonicalPath":"/agent/activepieces-activepieces","slug":"activepieces-activepieces","name":"activepieces","description":"AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents","url":"https://github.com/activepieces/activepieces","homepage":"https://www.activepieces.com","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-15T02:22:12.426Z","createdAt":"2026-02-25T03:38:12.412Z","downloads":null},{"id":"5cb26759-3a39-483f-94cf-276a98c13bb8","entityType":"agent","canonicalPath":"/agent/cherryhq-cherry-studio","slug":"cherryhq-cherry-studio","name":"cherry-studio","description":"AI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs","url":"https://github.com/CherryHQ/cherry-studio","homepage":"https://cherry-ai.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-11T14:38:40.986Z","createdAt":"2026-02-25T03:38:19.379Z","downloads":null},{"id":"6f6582d0-5d76-4f0f-b81d-86520247950b","entityType":"agent","canonicalPath":"/agent/copilotkit-copilotkit","slug":"copilotkit-copilotkit","name":"CopilotKit","description":"The Frontend for Agents & Generative UI. React + Angular","url":"https://github.com/CopilotKit/CopilotKit","homepage":"https://docs.copilotkit.ai","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-03-25T09:50:57.846Z","createdAt":"2026-02-25T03:39:14.617Z","downloads":null}],"links":{"hub":"/agent","source":"/agent/source/clawhub","protocols":[{"label":"OpenClaw","href":"/agent/protocol/openclew"}]}}}