Check loop.qnt's properties against the real loop (M2.1, PR D part 1)

test_loop_properties.py walks the new loop with a seeded mix of operator
and Gitea actions (go-aheads, stale go-aheads, answers, comments, merges,
the comment-then-merge race, closes, abandon, a crash and restart) and an
adaptive scripted model, checking every loop.qnt property after every
step. It found two bugs, both fixed with regression tests: abandoning an
item left its open PR open (the model's fix 1, now in code), and after a
review fix became a follow-up PR, the next fix went to the merged branch.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PPQ14L3LikyefCMFBqEbKE

nisse-dev/src/nisse_dev/loop/milestones.py

@@ -32,6 +32,7 @@ class Fixed:
     kind: Literal["pushed", "nothing", "new-pr", "closed", "failed"]
     note: str
     pr: PullRequest | None = None
+    branch: str = ""  # a new PR's branch, which later fixes go to
 
 
 class Milestones:
@@ -177,7 +178,7 @@ class Milestones:
                     summary=f"PR #{pr} merged before this review fix landed.",
                     issue=0,
                 )
-                return Fixed("new-pr", new.note, new.pr)
+                return Fixed("new-pr", new.note, new.pr, follow)
             if state.state != "open":
                 return Fixed("closed", f"PR #{pr} was closed", state)
             return Fixed("pushed", f"pushed {sha[:8]}", state)