{"slug":"docs/canonical/09-CAPABILITY-FABRIC-AND-ADK","title":"09 — Capability Fabric and ADK","section":"docs","sectionLabel":"Docs","status":"unclassified","type":"document","path":"docs/canonical/09-CAPABILITY-FABRIC-AND-ADK.md","bytes":1632,"description":null,"updated":null,"source":"integratewise/Integratewise-docs-wiki","content":"\n**Canonical reference:** `00-SPINEWORKSPACE-INTEGRATEWISE-CANON.md` §18  \n**Status:** Canonical\n\n---\n\n## 1. Capability Fabric\n\nCapability Fabric is the provider-neutral execution language between intent and implementation.\n\nA user or Twin should think in business capabilities such as `prepare_renewal`, `schedule_review`, `update_project_state` or `send_followup`, not in raw provider endpoints.\n\n## 2. Capability resolution routes\n\nCapability Fabric resolves how authorized intent can be performed through route families including:\n\n- tool-to-tool\n- memory fetch\n- MCP\n- agent-to-agent\n- API connector\n- API wrapper/function\n- local execution\n- AI execution\n- browser, workflow or human execution where needed\n\n## 3. Capability contract\n\nEvery executable capability declares: version, input/output schema, authority scope, risk, execution mode, sync mode, idempotency, evidence requirements, provider routes, health, timeout, retry, compensation and verification behavior.\n\n## 4. Responsibility split\n\n```text\nTwin / Workbench      = what the user is trying to achieve\nCapability Discovery  = what executable options exist\nCapability Fabric     = how it could be done\nGovernance            = whether it may be done and under whose authority\nHermes                = how approved execution is coordinated reliably\nReconciliation        = whether reality actually changed as intended\n```\n\n## 5. Capability Discovery\n\nDiscovery identifies available capabilities from: connectors, MCP servers, agents, workflows, APIs and platform services. Discovery is read-only; it does not authorize execution.\n"}