{"slug":"docs/canonical/10-GOVERNANCE-AUTHORITY-AND-APPROVAL","title":"10 — Governance, Authority and Approval","section":"docs","sectionLabel":"Docs","status":"unclassified","type":"document","path":"docs/canonical/10-GOVERNANCE-AUTHORITY-AND-APPROVAL.md","bytes":1024,"description":null,"updated":null,"source":"integratewise/Integratewise-docs-wiki","content":"\n**Canonical reference:** `00-SPINEWORKSPACE-INTEGRATEWISE-CANON.md` §19  \n**Status:** Canonical\n\n---\n\n## 1. Governance\n\nGovernance is a cross-cutting authority boundary, not merely a single UI screen or one step in a linear workflow.\n\nGovernance evaluates: who or what requested the action; tenant and organizational scope; human or agent role; delegated authority; capability and target; risk and destructive potential; data sensitivity; approval policy; evidence requirements; audit obligations.\n\n## 2. Canonical outcomes\n\n```text\nALLOW\nDENY\nREQUIRE APPROVAL\nDELEGATE / ALLOW WITH BOUNDED AUTHORITY\n```\n\n## 3. Human governance\n\nHuman governance does not mean a human must approve every low-risk automated operation. It means the organization defines what AI may do automatically, what requires explicit approval and what may never be delegated.\n\n## 4. Non-negotiable rule\n\nTwin cannot self-approve consequential action or bypass Governance. Confidence never grants authority.\n"}