AI agents can send emails, make purchases, post on social media, and control your home. Agent Constitution lets you decide what they can and can't do — with approval from your iPhone for every restricted action.
How It Works
Define rules on your iPhone. When your AI agent attempts a restricted action, you get a push notification. Review the action, approve or deny, and the agent proceeds only with your permission.
Set behavioral rules on your iPhone — what actions require your approval, across 6 categories.
Your AI agent attempts an action — sending an email, making a purchase, deploying code.
The action matches a rule. A push notification is sent to your iPhone with full details.
Review what the agent wants to do. Approve or deny. The agent proceeds only if you say yes.
Rules · Approval · Audit Trail
Stateless HTTP · HMAC Signed
Any agent · Any framework
Rule Categories
Pre-built templates for the most common agent actions, plus custom rules for anything else. Define your boundaries in plain English.
Features
Agent Constitution puts you in control — define your own rules, get notified in real time, and keep a complete audit trail of every decision.
You define the behavioral boundaries — not the AI company. 15 prebuilt templates across 6 categories, plus custom rules in plain English.
When a rule is triggered, you get an instant push notification on your iPhone. See exactly what the agent wants to do, then approve or deny.
Every rule check, every approval, every denial — logged on-device in the iOS Keychain. Complete history of what your agent did and didn't do.
Describe your boundaries in natural language and get intelligent rule suggestions. "Don't let my agent post anything political" → rules generated.
No SDKs, no vendor lock-in. Any AI agent framework can check rules with a single HTTP call. Works from Python, Node, bash, or any language.
The relay infrastructure is fully open source. Self-host, audit, fork. Your rules, your infrastructure, your control.
Audit Trail
Agent Constitution keeps a complete, tamper-proof audit log of every rule check, approval, and denial — stored securely on your device.
What the agent attempted — full action description, category, and timestamp
Whether you approved or denied, and which rule was triggered
Audit logs stored in iOS Keychain — encrypted, on-device, not synced to cloud
Filter by category, date, or decision. See patterns in your agent's behavior
For Developers
Your agent checks the constitution before performing restricted actions. If a rule applies, the user gets a push notification and the agent waits for approval. Standard HTTP — no SDK required.
📦 Install the agent skill: clawdhub install constitution
# Check the constitution before acting curl -X POST \ -H "Authorization: Bearer $SECRET" \ -d '{ "type": "constitution_check", "action": "send_email", "category": "communication", "description": "Send quarterly report to team@company.com", "details": { "to": "team@company.com", "subject": "Q4 Report" } }' \ "$RELAY/channel/$CH/push?role=gateway" # User gets push notification... # They review and approve/deny... # Pull the decision curl -H "Authorization: Bearer $SECRET" \ "$RELAY/channel/$CH/pull?role=gateway&wait=30" # → {"decision": "approved", "rule": "work_emails"}
Download Agent Constitution, define your boundaries, and take control
of what your AI agent can and can't do.