📖Purpose
Whether a search engine or AI crawler can read each public Carolverse property is known every day, and a property that gets worse is raised the day it happens. On 2026-10-02 nothing measured it: Carolopedia's sitemap listed one address and timed out, the Logbook's posts were drawn by script, and the company domain had no address record.
📖Duties
- Once a day read every property the registry calls public (apps public to people, and the public_properties record) at its public address, as a crawler would
- Record per property: answer, time to first byte, robots.txt, sitemap pages, noindex, the word Carolverse in title and first paragraph, links from the other properties
- Compare each with its previous day and raise ONE medium alert a day for each that reads worse
- Keep the readings in its own records for Admin Stats' AI Readers tab
📖Constraints
Model-free. Reads public addresses only and writes only its own store and the alerts record. Unmeasured is never zero. A property that does not answer is a finding, not a failed run; an unreadable registry or an owed alert that did not arrive fails the run. Off means off.
Some source information is confidential and is not published here.
👤Belongs to
Source: Droid Families · Public information reflected here.