fix(upstream-triage): distinguish subject hints from patch matches (#569)

* fix(upstream-triage): require patch identity before skipping commits

* fix(upstream-triage): retain subject-only ports as verification hints

* changelog: place the #569 entry under [Unreleased]

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Mads Lorentzen <madslorentzen_17@hotmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

CHANGELOG.md

@@ -426,6 +426,8 @@ per-file diff commands.
   most cards and `height:0` alone is a spacer. Three new `Bypasses` cases, seven subtests
   failing on master, plus a calibration case pinning that either half alone stays visible.
 
+- **Upstream triage distinguishes subject-only matches from verified patch matches** (#569) - matching subjects with different patches now name the fork commit and request verification instead of claiming the change was already cherry-picked.
+
 ### Security
 
 - **CI no longer leaves the job token in `.git/config` or expands a PR-controlled name into a

tests/test_upstream_triage.py

@@ -205,6 +205,61 @@ class AlreadyAppliedTests(TriageRepoFixture):
         self.assertIn("**1** worth reviewing", result.stdout)
 
 
+class SubjectCollisionTests(TriageRepoFixture):
+    def test_different_patch_with_same_subject_requests_verification(self):
+        self.write("kept.py", "print('upstream fix')\n")
+        upstream_sha = self.commit("Fix configuration")
+        self.set_upstream_to_head()
+        git(self.root, "reset", "--hard", "HEAD~1")
+        self.write("shared.txt", "fork configuration\n")
+        fork_sha = self.commit("Fix configuration")
+
+        result = self.run_triage()
+        self.assertEqual(result.returncode, 0, result.stderr)
+        self.assertIn("**0** worth reviewing", result.stdout)
+        self.assertIn(f"subject matches fork commit {fork_sha[:9]}, patch differs: verify", result.stdout)
+        self.assertNotIn(f"git cherry-pick {upstream_sha}", result.stdout)
+        self.assertNotIn("already applied (cherry-picked)", result.stdout)
+
+    def test_conflict_resolved_personalized_port_has_distinct_verify_reason(self):
+        profile = ".claude/skills/job-application-assistant/01-candidate-profile.md"
+        self.write(profile, "Name: template\nGuidance: old\n")
+        self.commit("Add template profile")
+        self.write(profile, "Name: template\nGuidance: improved\n")
+        upstream_sha = self.commit("Improve profile guidance")
+        self.set_upstream_to_head()
+        git(self.root, "reset", "--hard", "HEAD~1")
+        self.write(profile, "Name: personal\nGuidance: old\n")
+        self.commit("Personalize profile")
+        picked = subprocess.run(["git", "cherry-pick", upstream_sha], cwd=self.root,
+                                capture_output=True, text=True)
+        self.assertNotEqual(picked.returncode, 0, "fixture must exercise a real conflict")
+        self.write(profile, "Name: personal\nGuidance: improved\n")
+        git(self.root, "add", profile)
+        git(self.root, "cherry-pick", "--continue")
+        fork_sha = git(self.root, "rev-parse", "HEAD").strip()
+        result = self.run_triage()
+        self.assertEqual(result.returncode, 0, result.stderr)
+        self.assertIn("**0** worth reviewing", result.stdout)
+        self.assertIn(f"subject matches fork commit {fork_sha[:9]}, patch differs: verify", result.stdout)
+        self.assertNotIn("already applied (cherry-picked)", result.stdout)
+
+    def test_same_patch_with_renamed_subject_is_still_already_applied(self):
+        self.write("kept.py", "print('upstream fix')\n")
+        upstream_sha = self.commit("Fix configuration")
+        self.set_upstream_to_head()
+        git(self.root, "reset", "--hard", "HEAD~1")
+        self.write("fork_only.txt", "fork\n")
+        self.commit("Fork configuration")
+        git(self.root, "cherry-pick", upstream_sha)
+        git(self.root, "commit", "--amend", "-m", "Port upstream configuration fix")
+
+        result = self.run_triage()
+        self.assertEqual(result.returncode, 0, result.stderr)
+        self.assertIn("**0** worth reviewing", result.stdout)
+        self.assertIn("already applied (cherry-picked)", result.stdout)
+
+
 class SuggestedCherryPickOrderTests(TriageRepoFixture):
     def test_suggested_commands_apply_dependent_changes_in_ancestry_order(self):
         self.write("kept.py", "print('first upstream change')\n")

tools/upstream_triage.py

@@ -18,8 +18,10 @@ Two signals drive the sort:
 
 1. Already applied? A cherry-pick lands with a NEW sha but the same patch, so a
    raw sha comparison misreports it as missing. We compute git patch-ids for the
-   fork-only commits and treat any upstream commit whose patch-id (or exact
+   fork-only commits and treat any upstream commit whose patch-id
-   subject) matches as already applied.
+   matches as already applied. A matching subject with a different patch may be a
+   conflict-resolved port or an unrelated change: keep it as a distinct verify
+   reason in the advisory skip list, never as proof of a cherry-pick.
 
 2. Relevant to this fork? A commit that only touches files this fork deleted
    (e.g. removed demo portals) is almost certainly N/A. We check each commit's
@@ -175,7 +177,9 @@ def main() -> int:
 
     fork_only = rev_list(f"{ref}..HEAD")
     fork_patch_ids = {p for p in (patch_id(s) for s in fork_only) if p}
-    fork_subjects = {subject(s) for s in fork_only}
+    fork_subjects = {}
+    for fork_sha in fork_only:
+        fork_subjects.setdefault(subject(fork_sha), fork_sha)
     try:
         common_base = git("merge-base", "HEAD", ref).strip()
     except subprocess.CalledProcessError:
@@ -188,9 +192,14 @@ def main() -> int:
     for sha in behind:
         subj = subject(sha)
         short = sha[:9]
-        if patch_id(sha) in fork_patch_ids or subj in fork_subjects:
+        if patch_id(sha) in fork_patch_ids:
             skip.append((short, sha, subj, "already applied (cherry-picked)"))
             continue
+        if subj in fork_subjects:
+            match = fork_subjects[subj][:9]
+            skip.append((short, sha, subj,
+                         f"subject matches fork commit {match}, patch differs: verify"))
+            continue
         if any(sha.startswith(e) for e in wontport):
             skip.append((short, sha, subj, "on the fork's won't-port list"))
             continue