Two layers, checked in order
- Reachability — can the calling user reach the connector at all? They can if they own it, it’s shared with them (by username, or on Pro and Enterprise with an org unit they belong to — see below), or it’s a platform-managed integration they’ve connected their own account to. A connector that fails this check is invisible — it doesn’t appear in the tool list.
- Per-agent permission — is the connector enabled for this agent, and is the specific tool allowed?
No configuration means fully allowed
Per-agent permission is default-allow. Sharing a connector with someone makes it usable on every agent they run, immediately. A permission row exists only to restrict: disable a connector for one agent, or apply a rule to its tools.
Rules are glob patterns against tool names:
SEND_ tool is blocked, reads are allowed, and deletes need approval.
Configuring it
Example body for the
PUT:
Sharing a connector with an org unit
Plans: Pro and Enterprise. Self-hosted: Enterprise license.
admin can also share it with an org unit. A unit grant reaches everyone in that unit and in every unit beneath it.
Share from the connector’s Access tab in the dashboard, or with the API:
Revoking a share also deletes the grantee’s stored credential for that connector. A later re-grant starts from a clean reconnect.
Related
- MCP overview — connectors, connections, request flow
- User → agent access
- Access control overview