Wiki Generation Architecture: Multi-Agent Multi-Pass Design

{
  "id": "wiki-generation-architecture",
  "wikiId": "E766",
  "path": "/internal/wiki-generation-architecture/",
  "filePath": "internal/wiki-generation-architecture.mdx",
  "title": "Wiki Generation Architecture: Multi-Agent Multi-Pass Design",
  "quality": 55,
  "readerImportance": 75,
  "researchImportance": null,
  "tacticalValue": null,
  "contentFormat": "article",
  "causalLevel": null,
  "lastUpdated": "2026-03-23",
  "dateCreated": "2026-02-16",
  "summary": "Architecture proposal for next-generation wiki page generation. Analyzes current single-pipeline limitations, surveys state-of-the-art (Stanford STORM, GraphRAG, CrewAI, Self-Refine, Loki/OpenFactVerification, GPT-Researcher, RefChecker), and proposes a multi-agent multi-pass architecture with 8 specialist agents, 12+ composable passes, knowledge graph-driven linking, and dynamic computation embedding. Includes concrete implementation plan, composable module architecture, and academic foundations drawing on proposition-level retrieval (Dense X Retrieval), FActScore verification, nanopublications, and argumentation frameworks.",
  "description": "Architecture proposal for scalable, high-quality wiki page generation using specialist agents, multi-pass refinement, knowledge graph integration, and dynamic computation",
  "ratings": {
    "novelty": 7,
    "rigor": 6,
    "completeness": 7,
    "actionability": 8
  },
  "category": "internal",
  "subcategory": "architecture",
  "clusters": [
    "ai-safety"
  ],
  "metrics": {
    "wordCount": 5288,
    "tableCount": 11,
    "diagramCount": 3,
    "internalLinks": 2,
    "externalLinks": 16,
    "footnoteCount": 10,
    "bulletRatio": 0.25,
    "sectionCount": 68,
    "hasOverview": false,
    "structuralScore": 13
  },
  "suggestedQuality": 87,
  "updateFrequency": null,
  "evergreen": false,
  "wordCount": 5288,
  "unconvertedLinks": [
    {
      "text": "Anthropic's own research system",
      "url": "https://www.anthropic.com/engineering/multi-agent-research-system",
      "resourceId": "b82761b098942cc5",
      "resourceTitle": "How we built our multi-agent research system \\ Anthropic"
    }
  ],
  "unconvertedLinkCount": 1,
  "convertedLinkCount": 0,
  "backlinkCount": 1,
  "hallucinationRisk": {
    "level": "medium",
    "score": 50,
    "factors": [
      "low-citation-density"
    ]
  },
  "entityType": "internal",
  "redundancy": {
    "maxSimilarity": 15,
    "similarPages": [
      {
        "id": "language-models",
        "title": "Large Language Models",
        "path": "/knowledge-base/capabilities/language-models/",
        "similarity": 15
      },
      {
        "id": "architecture",
        "title": "System Architecture",
        "path": "/internal/architecture/",
        "similarity": 15
      },
      {
        "id": "knowledge-graph-ontology",
        "title": "Knowledge Graph Ontology: Design & Implementation Status",
        "path": "/internal/knowledge-graph-ontology/",
        "similarity": 15
      },
      {
        "id": "agentic-ai",
        "title": "Agentic AI",
        "path": "/knowledge-base/capabilities/agentic-ai/",
        "similarity": 14
      },
      {
        "id": "solutions",
        "title": "AI Safety Solution Cruxes",
        "path": "/knowledge-base/cruxes/solutions/",
        "similarity": 14
      }
    ]
  },
  "changeHistory": [
    {
      "date": "2026-02-17",
      "branch": "claude/wiki-generation-architecture-0J97Q",
      "title": "Route internal pages through /wiki/E<id>",
      "summary": "Migrated internal pages from `/internal/` to `/wiki/E<id>` URLs so they render with full wiki infrastructure (breadcrumbs, metadata, quality indicators, sidebar). Internal MDX pages now redirect from `/internal/slug` to `/wiki/E<id>`, while React dashboard pages (suggested-pages, updates, page-changes, etc.) remain at `/internal/`. Follow-up review: cleaned up dead code, hid wiki-specific UI on internal pages, fixed breadcrumbs, updated all bare-text `/internal/` references.",
      "pr": 182
    },
    {
      "date": "2026-02-16",
      "branch": "claude/wiki-generation-architecture-DYIaD",
      "title": "Wiki generation architecture research & proposal",
      "summary": "Researched state-of-the-art approaches to scalable wiki generation (Stanford STORM, Microsoft GraphRAG, CrewAI, Self-Refine, SemanticCite, Anthropic multi-agent systems, KARMA) and wrote a comprehensive architecture proposal for multi-agent, multi-pass wiki page generation. The proposal covers 8 specialist agents, 12+ composable passes, knowledge graph-driven content planning, dynamic computation embedding, and iterative refinement loops.",
      "pr": 173
    }
  ],
  "coverage": {
    "passing": 3,
    "total": 13,
    "targets": {
      "tables": 21,
      "diagrams": 2,
      "internalLinks": 42,
      "externalLinks": 26,
      "footnotes": 16,
      "references": 16
    },
    "actuals": {
      "tables": 11,
      "diagrams": 3,
      "internalLinks": 2,
      "externalLinks": 16,
      "footnotes": 10,
      "references": 1,
      "quotesWithQuotes": 0,
      "quotesTotal": 0,
      "accuracyChecked": 0,
      "accuracyTotal": 0
    },
    "items": {
      "summary": "green",
      "schedule": "red",
      "entity": "red",
      "editHistory": "green",
      "overview": "red",
      "tables": "amber",
      "diagrams": "green",
      "internalLinks": "amber",
      "externalLinks": "amber",
      "footnotes": "amber",
      "references": "amber",
      "quotes": "red",
      "accuracy": "red"
    },
    "editHistoryCount": 2,
    "ratingsString": "N:7 R:6 A:8 C:7"
  },
  "readerRank": 124,
  "recommendedScore": 152.86
}

No external links