Quarterly retest schedule
Keeps the date each reviewed agent was last tested and when it is due again, with a short set of core tasks, and raises an alert when a retest is overdue.
- Status
- Live
- Last checked
- not yet
- Agents on it
- 1
What the quarterly retest schedule is
Agents change faster than their reviews. A model update can fix the failures a review reports or break what worked. The schedule makes sure every reviewed agent is tested again on the same core tasks a quarter after its last test, and that I cannot forget one.
The core tasks are T3 (Spreadsheet from six invoices), T8 (Fill in the 12-field booking form), T15 (Research 30 cafés at once), drawn from the full test pack so a retest is comparable with the first test.
Runs on: A schedule file checked by the daily monitor health check. Manus: fresh, next retest due 26 January 2027.
How it works
- Each reviewed agent has a last-tested date and a due date 90 days later. When a retest is done, the script records the date and the version tested and sets the next due date.
- Every day the monitor health check recomputes each agent's status: fresh, due soon when the due date is within two weeks, or overdue.
- An overdue agent becomes a problem in the health check, which opens an issue that reaches me; one due soon becomes a note.
- The review's tools card shows the last test date and the next due date, read from the schedule file.
What it records
- Per agent
- Last tested, next due, status, the product version tested and a history of every test with a note.
- Core tasks
- The task ids and titles a retest must repeat, and why each was chosen.
What it holds today
Read from the monitor’s data file when this page was built. Every reviewed agent is added when its review is made.
Manus
| Figure | Value |
|---|---|
| Last tested | 28 October 2026, Manus 2.0 Lite (free plan) |
| Next retest due | 26 January 2027 |
| Status | fresh |
Figure
- Last tested
- Value
- 28 October 2026, Manus 2.0 Lite (free plan)
- Next retest due
- Value
- 26 January 2027
- Status
- Value
- fresh
What it cannot catch
- It schedules and alerts. The retest itself is done by hand with the same test world and checkers as the first test.
- The status on the site is the one stored at the last run of the script or the health check, so it can lag a day.
Where it appears on this site
The other AI agent monitors
- Price and limits tracker
- Product uptime
- Terms and privacy change watcher
- Trustpilot tracker
- Support response timer
- All monitors
Written by Miles Foster, Owner and tester, RankDude. How RankDude tests and grades AI agents is on the AI agents methodology page.