fix(eval): raise iteration-followup run cap to 25

The agent legitimately runs 20+ diagnostic commands for a config_mutation
task (reset service, re-run backup, verify, check logs). Cap of 8 was
too strict.

evals/iteration-followup.yaml

@@ -15,4 +15,4 @@
     - kind: proposes_plan
     - kind: writes_back
     - kind: max_run_calls
-      value: 8
+      value: 25