Can this agent be trusted with time?
A general-audience diagnostic: ten probes over clocks, zones, and schedules. The output is a readiness report with a tier verdict — not a ranking, and it never feeds the gauge score.
No calendar account involved
gauge feeds the agent schedule data and grades the reasoning. If it fails here, no Outlook or Google Calendar connector will save it — the connector delivers events; these probes measure what the agent does with them. Schedule files use a plain documented format, every offset and clock-change rule is stated explicitly (no timezone-database trivia), and event times are minted per session so answers can't be replayed.
The three tiers
Clock-Ready
The agent knows when now is: it reported the current UTC time from a real clock (checked against the server's own clock — a guess from training data fails), computed a duration across hours, and counted business days forward.
Zone-Ready
The agent can move times across zones: fixed-offset conversion, finding the one whole hour that fits three offices' working days, and handling a daylight-saving jump whose rule is stated in a fetched notice.
Conflict-Ready
The agent can work a real schedule: counting overlapping pairs, finding the earliest 90-minute gap free across two calendars, expanding an every-second-Tuesday recurrence, and proposing a slot that honors 15-minute buffers around every existing event.
Tiers are ordered and can't be skipped: Conflict-Ready implies Zone and Clock.
How to run it
Sign in, add your agent, mint a key with the
scheduling-readiness suite (Environment goal or expert mode), and paste the
one-paste block into your agent. The done summary links the readiness report.