Scheduled Agent
Scheduled Agent
Also called: 定时智能体 · Cron agent
An agent that runs automatically on a fixed cadence such as cron, with no human trigger — used for monitoring, scraping, digests, and periodic checks.
Typical scenarios
Summarise yesterday's metrics into a morning brief; check a competitor's page hourly for changes; sweep dependencies weekly for security updates; process the day's backlog queue overnight. The common thread: the task repeats on a schedule, can run unattended, and produces a report or a follow-up action. It is a specialised form of the time-based trigger.
What unattended operation demands
With no one watching, it depends more on: observability (a full trace per run, failures traceable), idempotency (re-running the same period causes no duplicate side effects), alerting (proactively notify a human on error), and an explicit spend cap. Otherwise a quietly failing or runaway scheduled job can go unnoticed for a long time.
Common misconceptions
- A scheduled agent is not a "simpler" agent — unattended operation raises the bar for error handling and observability.
- Hanging an interactive agent on cron without idempotency and alerting accumulates silent failures.
Related terms
Sources
Compiled 2026-08-29 · This glossary is compiled from public papers, official specifications, and common industry definitions, and is updated as the field evolves. Corrections welcome.