These pages are implementation scenarios, not unverifiable customer proof claims. Use them to understand how Stage0 fits into real high-risk workflows.
Payouts, account changes, and external notifications
Use Stage0 to separate low-risk automation from actions that need approval, role checks, or environment-aware blocking.
AI-assisted deploys, migrations, and rollback readiness
Use Stage0 to gate production-changing actions with environment rules, change windows, and approval evidence before any tool runs.
MCP servers exposing internal tools
OpenAI Agents SDK or LangGraph tool-use loops
Deployment and database migration approval flows
Payments, deletes, and external notification workflows
What happens when Stage0 times out or returns 5xx?
Which fields get logged for audit and incident review?
How do approvals, roles, and environment scope map into `/check`?
Who owns the final enforcement step in production?
Start with the docs for quickstarts, fail-closed behavior, quota rules, audit fields, and runtime contract details.