Your AI agent can run your content operation. This pack is what keeps the judgment yours.Three governed workflows for the agent you already use: a content audit, a content brief system, and an editorial pipeline. The agent reads, scores, drafts, runs the checklists, and tracks status. The decisions, approvals, and publish calls are reserved for humans, and the reservation is written into every procedure and every artifact.What makes it governed, not just prompted: A permission map for every field in every workflow Human fields stay visibly empty, marked "awaiting human" A human may dictate a decision; the agent may never make one. No bypass phrase exists, and every override is recorded in the artifact itself Provenance on everything: what was read, what was suggested, who decided Inside: the boundary document, eight core procedure files (scoring rubrics with exact scale points, two twelve-item binary checklists, permission maps, the provenance format), five platform adapters with paste-ready instructions and smoke tests (Claude Projects, Claude Code, Custom GPT, Microsoft 365 Copilot, MCP agents), a 20-minute first-audit guide, and a plain-language license.Tested before shipping. A fresh agent given only this pack ran a real audit, refused to fill the decision fields when asked, and recorded the request in the provenance. Twice, adversarially.Works standalone. Pairs with The Fully Inhabit Suite, the same workflows as structured Notion systems where the boundary lives in the database schema. One purchase covers you and your own team.