- tags
- #help #security #inherited #wip
- order
- 75
- description
- your AI's tokens, the vault and what your AI may do on its own
- brief
- Security is where you decide what your AI can reach: its tokens, the secrets in your vault, and what it may do on its own or must ask you first.
Help: Security¶
The Security pages (user menu → Settings → AI permissions) are where you decide what your AI can reach and do on a project. They are three tabs:
- AI tokens (/ai/tokens): a token lets an AI, or a program such as a sync plugin, act as you on this project, at most at the level you pick and never above your own role. Make one, give it to your AI, renew it or end it here. The page's own steps show how to give it to your AI.
- Vault (/ai/vault): secrets your agents need (a key for another service). Only you and the agents you allow can use them; a value is never shown again after saving. More: Vault.
- Permissions (/ai/permissions): for every action, whether your AI does it on its own, tells you, asks you first, or leaves it to a person. More: Permissions.