1. OCX on H1 acts on R1. OCX on H2 does nothing with it because R1's `clone.user` is "lab-1" and the clone at H2 also has server-code "aaaa…a" and its clone's user is "lab-1", which differs from R1.

2. No, OCX on H3 does not act on R1 because the run was sent by a different sender ("stone") than lab-9, who is named in the `clone` field of R1. If H3's clone's user were "lab-1", OCX would still act on it for the same reason.

3. Yes, mallory's `ended` counts because it satisfies all the rules for an `ended`: it names the correct run (`R1`) and a sender that holds admin or setup on the server.

4. No, stone's `started` does not count because OCX checks if the run's clone has the same `user` as the one the run is acting as on the server, which in this case would be "builder". Stone cannot act as "builder".

5. OCX on H1 acts on R3 because it is sent by a user (stone) holding setup and meets all other requirements. Mallory's `started` does not count for the same reason as before.

6. OCX on H1 acts on R4, but mallory's `started` does not count because the run's process user "builder" works for another clone (the one at H2).

7. No, stone cannot send a run named with `"process": {"user": "lab-1"}` and have it act on the server because that would allow stone to act as lab-1, which they do not.

8. OCX on H2 acts on R5 because it is sent by a user (stone) holding setup and meets all other requirements.

9. The most OCX on H1 can do on the server is what stone holds, which includes check-in (`i`). Stone chose this when setting up their account.

10. `$$process_clone$$` is set to `$$workspace$$/clone`, and it is not the repository R1's key `clone` names because that refers to H1's clone, while process_clone is about creating a new clone on H2.

11. OCX on H1 acts on the new run, but mallory must send a new run with the updated server-code since the old one no longer matches any existing clone.

12. Only users holding admin or setup can send a `stop` for R1. A `stop` holds: `ocx`, `kind`, and `run`.