1Password
What Hal reads
Three feeds from each 1Password account, through its Events API:
- Sign-in attempts — every attempt to sign in to the account, with its outcome (success, credentials failed, MFA failed, blocked by a firewall rule) and the exact reason, the app and platform used, the address and location. In 1Password an MFA failure means the account password and Secret Key were right and the second factor was not; Hal reads it so.
- Item usages — who filled, revealed, copied, edited, exported or shared an item in a shared vault, by vault and item. Hal sees vault and item identifiers, never names or contents.
- Audit events — the account’s change log: recoveries, account password and Secret Key changes, MFA and device changes, vault exports and shares, service-account tokens, access grants, and, in your own account, your technicians’ launches into managed companies.
Each event lands in that client’s triage with everything else Hal sees. The first pull reaches back 30 days; after that Hal reads the feeds every cycle.
Access is read-only, and that is enforced by the credential, not by a promise: an Events Reporting token can read events and nothing else. Hal cannot read an item’s contents, cannot change anything in the account, and holds no service account.
One token per account
Every 1Password account issues its own token. For each managed company you want Hal to read, you issue the token inside that company’s account after Launching into it from your MSP console; for your own company you issue one signed in to your own account. Your own company is a client of your Hal like any other, and Hal knows it is yours: ask Hal to set your company up as a client and say that it is you.
A client can hold more than one account. A company that is part of a client and has its own account in your MSP console, a subsidiary under its own name, say, connects under that client: its token is issued in that company’s account, after a Launch into it. Each account Hal reads has its own token and its own source.
Before you start
- You need to be an owner or administrator of the account (for a managed company, your technician’s launched session is one).
- Note which region the account is hosted on:
1password.comfor most,1password.ca,1password.euorent.1password.comotherwise. Hal asks for it.
Issue the token
- From your MSP console open Managed Companies and Launch into the
company whose account you are connecting. It opens in a new tab on that
company’s own address, such as
acme.1password.com; note it, because Hal asks for it, and the rest happens there. (For your own company’s first account, sign in to your own account; its address is the one you sign in at.) - In the sidebar select Integrations. If other integrations are already listed, select Directory.
- In the Events Reporting section choose Other, name the
integration
Hal, and select Add Integration. - Token Name:
Hal. Expires After: leave Never; you can revoke the token from this page at any time, and if your policy requires an expiry, Hal warns you two weeks before it and you rotate it then. Events to Report: all three — sign-in attempts, item usages, audit events. - Select Issue Token. The token is shown once. 1Password offers to save it in a vault; that is fine.
Hand it to Hal
In the Hal portal chat, as a Hal admin, tell Hal which client this account belongs to: “Connect 1Password for Acme.” He opens a paste box for that client under Settings → 1Password (it stays open for a day; asking again re-opens it) and relays the steps above. Paste the token, pick the region, Save, and tell him you’re done. The token never passes through chat. For a client that already has an account, ask the same way, “Connect another 1Password account for Acme”: Hal lists the accounts it has, and his steps launch into the company whose account you are connecting.

When you say done, Hal reads the paste and checks it against 1Password: he
confirms which account issued it and which of the three event types it may
read, and because the API does not name the account, he reads who signed in
to it over the last week — how many people, their email domains, how many
are your own technicians — and the newest sign-in, and asks you to confirm
this is Acme’s account and to give its sign-in address, the address from
step 1, such as acme.1password.com. 1Password exposes no address for an
account, so Hal checks its form: one name before the 1Password suffix, the
suffix of the region the token works in (.1password.com,
.ent.1password.com, .1password.ca or .1password.eu), and that no other
1Password account on your Hal has it. He stores the name before the suffix,
acme, as the account’s name. Nothing is stored until you confirm. Say yes and he stores the account and the token, creates the
account’s own 1Password log source and connects it to the detection
pipeline; the first pull runs in his next cycle, 30 days back. Say no and he
closes the box and writes nothing. If an address is wrong later, ask Hal to
correct the account’s sign-in address; nothing else about it changes.
If the token is rejected, was issued without one of the three event types, names an event type that could not be read just then, or belongs to an account already connected — including your own account when you meant a company’s, which happens when the token was issued without Launching first — Hal says so and tells you what to change. To replace a token later, ask Hal to rotate Acme’s 1Password token: he opens the box again, and the account the new token belongs to is the one whose token he replaces, so the client’s other accounts keep theirs. A token of an account that is not Acme’s is refused, and nothing is replaced.
What Hal shows
Settings → 1Password lists the connected accounts as a JSON you can
generate on the page or download: client, account, region, the token’s
identity and expiry, the event types the token can read and when that was
last checked. The
Sources page shows each 1Password account as its own source with three
pills, one per feed, and its name, such as acme, in the Tenant column.
Alerts name the account beside its client where the client holds more than
one.
Related: the other integrations Hal reads, and how events move through triage.
See what Hal surfaces on your own clients.
No deck. Ask us anything first. When you want to see Hal on your own tenants, we sign a short evaluation agreement and stand up your instance; you connect one tenant, Hal watches it for 14 days, and we walk you through what he found.
- 01You ask us your questions. No deck, no demo dataset.
- 02You sign a short evaluation agreement, and we stand up your own instance.
- 03You connect one tenant from your own admin console. Hal watches it for 14 days.
- 04We walk through what he found. Keep going month to month, or revoke the scopes yourself and stop.