🛠️ How-to guide
<aside> 🔑
WHAT IT DOES
Credentials are the keys and logins your agents need to reach outside systems — a mapping service, a work-order system, a text-message provider. You store each one once for the whole workspace and refer to it by name, so no secret is ever typed into an agent or shown on screen again.
</aside>
Your Riverton 311 agent might look up a resident's ward in the city GIS service, open a work order for a pothole, and text a confirmation — each of those needs a key. Store them here once and every agent can use them safely. You can set a value, but you can never read it back, so a key can't leak out of the platform.
Every credential shows whether its value is set or still a placeholder, which agents refer to it, and when it was last set and last used. You refer to a credential from an agent's connection settings by writing its name as {{kv:name}} — for example {{kv:gis-portal-key}}.

The Credentials list — each credential's status, the agents that refer to it, and when it was last set and last used.

Select Add credential to create a placeholder, then set the secret value on the next step.
gis-portal-key. This creates a placeholder.{{kv:gis-portal-key}}. The agent fills in the real value only at the moment it makes the call.<aside> ⭐
TIP
Before deleting a credential, check its Referenced by count. One marked Unreferenced is safe to remove; deleting a credential that agents still use will make those calls start failing.
</aside>
Can I see a value after I save it?
No. Credentials are write-only by design — you can set or rotate a value, but never view it. This keeps secrets from leaking through the screen.
How does an agent use a credential?
You refer to it by name as {{kv:name}} in the agent's connection settings. The real value is filled in only at the moment of the call.