Permission Slip

Permission Slip is the approval layer for your AI agent. When your agent needs to take an action — sending emails, merging pull requests, executing transactions, or calling APIs — it asks you first. You review, approve or deny, and only then does it execute.

How it works:

Your AI agent submits a structured request. You get a push notification with a plain-language summary. You review the details, see the risk level, and tap Approve or Deny. Done.
Your credentials stay safe.

Permission Slip holds your API keys and passwords in an encrypted vault. Your agent never sees them — it only submits what it wants to do, not how. Zero credential exposure.

Key features:
∙ Real-time push notifications for new approval requests
∙ Risk assessment badges (Low / Medium / High) for every action
∙ Expiration timers so nothing lingers forever
∙ Full parameter review before you approve
∙ Biometric lock with Face ID, Touch ID, or fingerprint
∙ Approve or deny with haptic feedback — one tap
∙ History of all approved and denied requests
∙ Per-device notification controls

Built for developers and teams who run autonomous agents but need human oversight. Whether you’re automating DevOps, managing infrastructure, or letting AI handle routine tasks, Permission Slip keeps you in control without slowing you down.
Works with Openclaw and any agent that integrates with the Permission Slip API. Self-hostable for teams that need it.