Scheduled execution

Scheduled execution means CLEA runs a saved agent prompt from the calendar.

The prompt can be daily, weekly, monthly, or set to one date. It can also mention workflow files and other workspace files.

The prompt text tells CLEA what to do.

Keep it direct. A good scheduled prompt names the evidence and the output.

Mentioned files give CLEA more context.

For example, a scheduled prompt can mention a workflow file and a sitemap file. The workflow says how to work. The sitemap says where page work should happen.

CLEA can also use dashboard context such as Analytics, Sources, Competitors, Calendar, Search Console, and Google Analytics when available.

The useful output depends on the routine.

Some routines should send a suggestion card. Some should return a summary. Some should create or edit a file. Some should only report that no action is needed.

Use suggestion cards when the next action is clear.

Use reports when the user needs understanding before action. Do not turn every report into a card.

Goal: run weekly visibility review.
Inputs: Analytics, Sources, Competitors, @workflow 2.
Steps: find the strongest movement, inspect evidence, decide if action is safe.
Output: short report and one suggestion card only if useful.

If a scheduled run fails or returns weak output, do not keep retrying blindly.

Check whether the prompt is too broad, the mentioned file is unclear, or the needed data is missing.