docs(harbor): warn in the task instruction that baseline evals create no candidate

vero/src/vero/harbor/build/templates/instruction.md.j2

@@ -18,7 +18,10 @@ progress on the splits you *are* allowed to evaluate, within a fixed budget.
 4. Check budget / which splits are evaluable anytime: `vero harbor status`.
 {% if submit_enabled %}5. When done, nominate your best commit: `vero harbor submit`.{% else %}
 The best commit you evaluate on `{{ selection_split }}` is selected automatically and
-scored on the hidden test split at the end.{% endif %}
+scored on the hidden test split at the end. Only commits *other than the seeded
+baseline* are selectable: evaluating the unmodified baseline spends budget without
+creating a candidate, so make sure at least one eval is of a commit that contains
+your changes.{% endif %}
 
 ## Rules
 

vero/tests/test_harbor_build.py

@@ -167,3 +167,12 @@ def test_seed_documents_advisory_read_only(built):
     seed = (built / "environment/main/seed.sh").read_text()
     assert "ADVISORY ONLY" in seed
     assert "sidecar-side" in seed
+
+
+def test_instruction_warns_baseline_not_selectable(built):
+    # auto_best: the agent must be told baseline evals do not create candidates
+    # (found live: an optimizer that spent its whole budget measuring the
+    # baseline died with "no candidate experiments" at finalize).
+    text = (built / "instruction.md").read_text()
+    assert "other than the seeded" in text
+    assert "spends budget without" in text