<aside> 🗓️
A scored calendar of recurring SEO and website quality checks - with evidence for every run.
Know which checks to install first, how often to run each, how much an AI agent may do alone, and whether last week's run actually has proof behind it.
</aside>
Use the three fictional sample sites first. They show a pass, a run with findings, a run that could not finish, and one incomplete run with a red gate.
A run counts only when its Run Gate is green. That needs a task, a site, a date, a result, and an evidence link - plus a next action whenever the result is Findings. A check with no evidence did not happen.
Human Approved is ticked by a person. An AI agent may fill everything else and must leave that box alone.
Each task is scored 1-5 on four factors, weighted and scaled to 100: Impact 35%, Stack Fit 25%, Measurability 20%, Automation Safety 20%. Re-score Stack Fit for your own tools - it is the factor most likely to differ from the samples.
| Score | Band | What to do |
|---|---|---|
| 90-100 | 🟢 Run first | Install this week. These are high impact and safe to run often. |
| 80-89 | 🟡 Schedule | Add within the month, at the listed cadence. |
| Below 80 | ⚪ Batch | Review quarterly, or fold into a larger audit. |
| Status | Meaning | What to do |
|---|---|---|
| 🔴 Overdue | Installed, and the last run is older than the cadence allows | Run it today or pause it on purpose |
| 🟠 Never run | Installed, with no run logged | Log the first run |
| 🟢 On schedule | Last run is inside the cadence window | Nothing |
| ⚪ Event-driven | Runs on releases or changes, not on a clock | Run it when the event in "Also Run On" happens |
| ⚪ Not installed | Not on your calendar yet, or paused | Decide by Priority Band |