{
  "schemaVersion": 3,
  "id": "article:ai-delegation-orchestration-06-capability-contracts-for-agent-networks",
  "slug": "ai-delegation-orchestration-06-capability-contracts-for-agent-networks",
  "title": "Capability Contracts for Agent Networks",
  "canonicalPath": "/articles/ai-delegation-orchestration-06-capability-contracts-for-agent-networks/",
  "sourcePath": "content/articles/2026/ai-delegation-orchestration-06-capability-contracts-for-agent-networks/article.md",
  "agentBriefPath": "content/articles/2026/ai-delegation-orchestration-06-capability-contracts-for-agent-networks/agent.md",
  "thesis": "Agent systems should be organized around replaceable capabilities with explicit contracts, not around vague role prompts or artificial replicas of human organizations.",
  "status": "published",
  "maturity": "sprout",
  "publishedAt": "2026-06-28",
  "updatedAt": "2026-06-28",
  "audiences": [
    "builders",
    "researchers",
    "students",
    "agents"
  ],
  "topics": [
    "ai-agents",
    "agent-orchestration",
    "human-ai-interaction"
  ],
  "series": {
    "slug": "ai-delegation-orchestration",
    "title": "AI Delegation Orchestration",
    "season": "Season 1",
    "order": 6,
    "role": "chapter"
  },
  "claims": [
    {
      "id": "claim-001",
      "claim": "Agent systems should be organized around replaceable capabilities with explicit contracts, not around vague role prompts or artificial replicas of human organizations.",
      "confidence": "medium",
      "status": "argument",
      "verification": {
        "reviewedAt": "2026-06-28",
        "reviewer": "human-publication-approval",
        "status": "verified"
      },
      "evidence": [
        {
          "sourceId": "source-mcp-introduction",
          "snippet": "The Model Context Protocol frames external capabilities as standardized connections between AI applications and tools or data.",
          "supports": "direct",
          "assessedAt": "2026-06-28"
        },
        {
          "sourceId": "source-mcp-security-best-practices",
          "snippet": "MCP security guidance supports explicit authorization, trust, and boundary design for connected capabilities.",
          "supports": "direct",
          "assessedAt": "2026-06-28"
        },
        {
          "sourceId": "source-anthropic-building-effective-agents",
          "snippet": "Agent-building guidance supports composing tools, workflows, and agents instead of assuming one monolithic assistant.",
          "supports": "indirect",
          "assessedAt": "2026-06-28"
        },
        {
          "sourceId": "source-hollan-hutchins-kirsh-distributed-cognition",
          "snippet": "Distributed cognition provides an analogy for intelligence spread across people, artifacts, tools, and environment.",
          "supports": "analogous",
          "assessedAt": "2026-06-28"
        }
      ],
      "counterevidence": [
        {
          "summary": "The delegation record and next-best-control models are proposed design primitives, not accepted standards.",
          "assessedAt": "2026-06-28"
        }
      ]
    }
  ],
  "sources": [
    {
      "id": "source-mcp-introduction",
      "title": "Model Context Protocol introduction",
      "url": "https://modelcontextprotocol.io/docs/getting-started/intro",
      "type": "protocol",
      "accessed": "2026-06-28"
    },
    {
      "id": "source-mcp-security-best-practices",
      "title": "Model Context Protocol security best practices",
      "url": "https://modelcontextprotocol.io/docs/tutorials/security/security_best_practices",
      "type": "documentation",
      "accessed": "2026-06-28"
    },
    {
      "id": "source-anthropic-building-effective-agents",
      "title": "Anthropic, Building effective agents",
      "url": "https://www.anthropic.com/research/building-effective-agents",
      "type": "article",
      "accessed": "2026-06-28"
    },
    {
      "id": "source-hollan-hutchins-kirsh-distributed-cognition",
      "title": "Hollan, Hutchins, and Kirsh, Distributed Cognition",
      "url": "https://www.lri.fr/~mbl/Stanford/CS477/papers/DistributedCognition-TOCHI.pdf",
      "type": "paper",
      "accessed": "2026-06-28"
    }
  ],
  "related": [
    {
      "type": "topic",
      "id": "topic:ai-agents"
    },
    {
      "type": "article",
      "id": "article:agent-control-planes"
    }
  ],
  "agentInstructions": [
    "Use claim IDs as the retrieval unit.",
    "Treat maturity=sprout as an exploratory uncertainty marker.",
    "Do not present proposed orchestration patterns as accepted standards.",
    "Keep examples generic and detached from private implementations."
  ],
  "provenance": {
    "createdAt": "2026-06-28",
    "createdBy": "codex-agent",
    "agents": [
      {
        "role": "drafting-and-site-preview",
        "model": "gpt-5",
        "invokedAt": "2026-06-28",
        "inputHash": "sha256:5aa76f2040787afd8b777ea4ed4af82974b89a9552edb10d3a968e3853983d5c",
        "outputHash": "sha256:d62f207cb226a1921aa71632681502ea8a38ef984d5c0046334cff23ddbb8fa2"
      }
    ],
    "reviews": [
      {
        "reviewer": "sibling-agent",
        "reviewedAt": "2026-06-28",
        "status": "commented",
        "scope": [
          "series structure",
          "evidence anchoring",
          "privacy"
        ],
        "notes": "Earlier sibling review identified source-ledger gaps; those gaps were resolved before publication with article-specific source alignment.",
        "resolvedAt": "2026-06-28"
      },
      {
        "reviewer": "human",
        "reviewedAt": "2026-06-28",
        "status": "approved",
        "scope": [
          "publication",
          "reader experience",
          "privacy",
          "series structure"
        ],
        "notes": "Human maintainer approved the local preview for website publication after article layout, reading-flow, privacy, and series-structure review.",
        "contentHash": "e2828555014eaf339ec5e645ae23c648d51ce09019c7137833258056aa89b22f"
      },
      {
        "reviewer": "sibling-agent",
        "reviewedAt": "2026-06-28",
        "status": "approved",
        "scope": [
          "publication-gate",
          "source alignment",
          "provenance",
          "privacy",
          "generated artifacts"
        ],
        "notes": "Independent sibling review approved the final publication packet with no blocking findings after source-ledger, provenance, privacy, and generated-artifact checks.",
        "contentHash": "e2828555014eaf339ec5e645ae23c648d51ce09019c7137833258056aa89b22f"
      }
    ],
    "policy": {
      "id": "policy:default",
      "version": "1.0.0"
    }
  },
  "contentHash": "e2828555014eaf339ec5e645ae23c648d51ce09019c7137833258056aa89b22f",
  "generatedAt": "2026-06-29T00:00:00.000Z",
  "articleUrl": "https://aura-knowledge.github.io/articles/ai-delegation-orchestration-06-capability-contracts-for-agent-networks/",
  "agentJsonPath": "/agents/articles/ai-delegation-orchestration-06-capability-contracts-for-agent-networks.json",
  "agentMarkdownPath": "/agents/articles/ai-delegation-orchestration-06-capability-contracts-for-agent-networks.md",
  "sourceRepoPath": "content/articles/2026/ai-delegation-orchestration-06-capability-contracts-for-agent-networks/article.md",
  "sourceGitHubUrl": "https://github.com/aura-knowledge/aura-knowledge.github.io/blob/main/content/articles/2026/ai-delegation-orchestration-06-capability-contracts-for-agent-networks/article.md",
  "tokenEstimate": 176,
  "sectionOutline": []
}