Human Approval Gate Checklist
Approval rules for deciding when an AI agent can proceed and when a human must review first.
Always require human approval before an agent performs these actions:
- Sending external email, SMS, social posts, or client messages.
- Charging money, refunding, changing Stripe, or touching billing.
- Printing, exposing, moving, or rotating secrets and credentials.
- Publishing production changes with user, revenue, or client risk.
- Deleting data, contacts, issues, files, campaigns, or records.
- Making claims about customers, testimonials, results, or compliance.
Allow autonomous execution when all are true:
- The task is scoped to a reversible artifact, draft, branch, report, or local file.
- The agent has enough context to proceed without guessing.
- The final output can be reviewed before it affects customers or public channels.
- The agent reports exact receipts and blockers.
Get Feedback Fast
Use the template, then pressure-test it with other builders
The library is designed to shorten planning time. The community is there to challenge assumptions before you burn a sprint on the wrong scope.
Related Templates
Adjacent assets that usually get used together.
Template
Daily Operator Report Template
A concise daily report format for always-on AI agent command centers.
Template
Discord Command Room Setup Template
A Discord room structure for routing founder inputs, agent work, approvals, and daily reports.
Template
Founder Command Center Issue Template
A scoped GitHub Issue template for routing founder inputs into approval-gated AI agent work.
Read the Supporting Playbooks
Use the deeper blog posts when you need context, tradeoffs, or examples behind the template.
Blog
How to Build a Founder Command Center With AI Agents
A practical blueprint for turning Discord, GitHub Issues, Obsidian, cron jobs, and AI workers into a founder command center that ships work instead of losing context.
Blog
Systems as a Service: The New SaaS Model for AI Builders
AI builders should stop thinking only in dashboards and seats. Systems as a Service packages software, agents, QA, and human approval into repeatable outcome loops customers can buy.
Blog
What Is an Always-On AI Agent?
An always-on AI agent is not a chatbot. It is a scheduled, approval-gated operating loop that watches inputs, does bounded work, verifies results, and reports what changed.