Good vs. weak agent descriptions — side by side
3 min
Concrete before/after rewrites that show what the wizard wants to see.
Weak: "Send me daily updates." The wizard has no trigger time, no source, no idea what's in the update.
Better: "Every weekday at 8:30am, list my Outlook unread email from the last 24 hours, summarise the top 5 by subject, and email me the digest." Clear trigger (cron-shaped), concrete source (Outlook), concrete work (top 5, summarise), concrete output (email).
Weak: "Watch for new GitHub PRs." Watch how? On a schedule? On a webhook? Output where?
Better: "Every hour, list new GitHub PRs in org/repo opened in the last hour, classify each as bug/feature/chore, and post a summary to the #engineering Slack channel."
Weak: "Track my stock portfolio." No data source named, no schedule, no output.
Better: "Every weekday at market close, fetch quotes for AAPL, MSFT, NVDA from Polygon, compute the day's gain/loss, and email me a one-paragraph summary."
The pattern: TRIGGER + SOURCE + WORK + OUTPUT. If a part is missing, the wizard either asks you to fill in a placeholder later (slowing you down) OR makes a guess (often wrong).
Live recipes need the desktop
This article is a static preview. The in-app Help sidecar inside Avery NXR can fire each step against your live project — install the desktop to use it interactively.