GLPI + Memtro

Only when you ask, and only what it has been granted: create a ticket (no notification is sent); add a follow-up to a ticket, private by default (no notification is sent); change a ticket's title, description, status, urgency, impact, priority or category; delete a ticket, to the trash or purged for good; assign a technician, observer or requester to a ticket; assign a group to a ticket; solve a ticket with a solution (notifies the requester); add a task to a ticket, private by default (no notification is sent); create a problem (no notification is sent); create a change (no notification is sent); create an inventory item (computer, monitor, printer, phone, network device or peripheral); change an inventory item's name, serial, comment, status, location, manufacturer, user or technician; delete an inventory item, to the trash or purged for good. Every other tool reads, searches or lists. The Connections page labels the tools that act, an administrator grants each one by name to the teams that may use it, and a personal connection acts only after its owner allows it.

A personal connection serves only your own requests. A shared connection serves the organisation: everyone, or only the teams an admin chooses. Admins can always add shared connections; other members need the permission.

Credentials are encrypted at rest. Tool results are cached briefly so agent loops do not repeat identical calls, and the facts the model extracts are saved to your organisation's memory with provenance (system, record and link) so they can be checked and superseded.

Host and user token (remote access key). A GLPI administrator turns on Setup → General → API: Enable Rest API and Enable login with external token, and adds an API client whose IPv4 range includes Memtro's outbound addresses.

Yes. Turn events on in the connection's Events tab on the Connections page, then add the address it shows as a webhook in GLPI, choosing the same kinds; the tab shows GLPI's own steps. A job set to start "When an event arrives" then runs once for each matching event, with the event as its input. The run treats the event as outside data, never as instructions, and anything the run remembers is kept as unverified.