fix(ci,docs): use unittest discovery for scripts.tests and align the-walk default route

.claude/skills/visual-review-and-fix/references/the-walk.md

@@ -13,10 +13,10 @@ wording is wrong, do not edit one copy. See "If a promise is wrong" at the botto
 
 Three top-level views, and the fragment is the address:
 
-- `#n=sessions` is the default. Session operations leads because the exception-first Attention route
-  was built first and buried the reader's opening questions in live review (NUI-16 in
+- `#n=projects` is the default route. The v2 design moves the entry point up to Projects so the reader
+  first finds the project and then its sessions (NUI-16 in
   [`docs/design-next-ui.md`](../../../../docs/design-next-ui.md)).
-- `#n=projects` is the complete map.
+- `#n=sessions` is fleet operations.
 - `#n=attention` is the gate queue and the coverage disclosure. It has a nav entry as of #287, last
   of the three, and the `a` key and the reported-blocks chip in the header still reach it. That chip
   appears only while a block is reported, which is why it was never a nav entry's substitute.

.github/workflows/quality-gate.yml

@@ -202,20 +202,7 @@ jobs:
       - name: Run the full suite on the floor
         run: |
           python -m unittest discover -s cargento/skills/cargento/tests -t .
-          python -m unittest \
-            scripts.tests.test_validate_plugins \
-            scripts.tests.test_bump_version \
-            scripts.tests.test_lint_embedded \
-            scripts.tests.test_bench_collect \
-            scripts.tests.test_capture_hook \
-            scripts.tests.test_bench_event_latency \
-            scripts.tests.test_derive_prompt_shapes \
-            scripts.tests.test_capture_team_registry \
-            scripts.tests.test_capture_terminal_identity \
-            scripts.tests.test_capture_focus_raise \
-            scripts.tests.test_serve_operator_cockpit \
-            scripts.tests.test_mark_abstention \
-            scripts.tests.test_score_abstention
+          python -m unittest discover -s scripts/tests -t scripts/tests
 
   test:
     name: Tests + coverage threshold
@@ -242,20 +229,7 @@ jobs:
         run: |
           coverage erase
           coverage run -m unittest discover -s cargento/skills/cargento/tests -t .
-          coverage run -a -m unittest \
-            scripts.tests.test_validate_plugins \
-            scripts.tests.test_bump_version \
-            scripts.tests.test_lint_embedded \
-            scripts.tests.test_bench_collect \
-            scripts.tests.test_capture_hook \
-            scripts.tests.test_bench_event_latency \
-            scripts.tests.test_derive_prompt_shapes \
-            scripts.tests.test_capture_team_registry \
-            scripts.tests.test_capture_terminal_identity \
-            scripts.tests.test_capture_focus_raise \
-            scripts.tests.test_serve_operator_cockpit \
-            scripts.tests.test_mark_abstention \
-            scripts.tests.test_score_abstention
+          coverage run -a -m unittest discover -s scripts/tests -t scripts/tests
 
       # No `coverage report` here. It reads `fail_under` from pyproject and
       # exits non-zero, and it ran BEFORE the label check below, so a PR
@@ -388,12 +362,7 @@ jobs:
       - name: Run dashboard test discovery
         run: python -m unittest discover -s cargento/skills/cargento/tests -t .
       - name: Run script unit tests
-        run: >-
-          python -m unittest scripts.tests.test_validate_plugins
-          scripts.tests.test_bump_version scripts.tests.test_lint_embedded
-          scripts.tests.test_bench_collect scripts.tests.test_capture_hook
-          scripts.tests.test_capture_terminal_identity
-          scripts.tests.test_capture_focus_raise
+        run: python -m unittest discover -s scripts/tests -t scripts/tests
 
   # The single required status check. Branch protection requires this job, so
   # every job above must succeed (a skip or failure anywhere fails the gate).

AGENTS.md

@@ -218,15 +218,7 @@ git diff "$(git merge-base origin/main HEAD)"..HEAD \
   -- '*plugin.json' '*gemini-extension.json' | grep -E '^[+-].*"version"'
 coverage erase
 coverage run -m unittest discover -s cargento/skills/cargento/tests -t .
-coverage run -a -m unittest \
-  scripts.tests.test_validate_plugins scripts.tests.test_bump_version \
-  scripts.tests.test_lint_embedded scripts.tests.test_bench_collect \
-  scripts.tests.test_capture_hook scripts.tests.test_bench_event_latency \
-  scripts.tests.test_derive_prompt_shapes scripts.tests.test_capture_team_registry \
-  scripts.tests.test_capture_terminal_identity \
-  scripts.tests.test_capture_focus_raise \
-  scripts.tests.test_serve_operator_cockpit \
-  scripts.tests.test_mark_abstention scripts.tests.test_score_abstention
+coverage run -a -m unittest discover -s scripts/tests -t scripts/tests
 coverage report   # enforces the fail_under threshold from pyproject.toml
 # `test_capture_terminal_identity` and `test_capture_focus_raise` exercise AppleScript
 # against Terminal.app, and this suite now sends nothing. It used to: measured on a macOS desk with Terminal

scripts/tests/test_mark_abstention.py

@@ -231,7 +231,8 @@ def test_the_temp_file_is_gone_and_the_mode_is_owner_only(self) -> None:
             mark_abstention._write(path, {"v": 2, "marks": {}})
             self.assertTrue(os.path.exists(path))
             self.assertFalse(os.path.exists(f"{path}.tmp"))
-            self.assertEqual(0o600, os.stat(path).st_mode & 0o777)
+            if os.name != "nt":
+                self.assertEqual(0o600, os.stat(path).st_mode & 0o777)
 
     def test_a_rewrite_replaces_rather_than_appends(self) -> None:
         with tempfile.TemporaryDirectory() as root:

scripts/tests/test_validate_plugins.py

@@ -1540,3 +1540,42 @@ def test_mutated_native_name_route_and_gate_mapping_are_rejected(self) -> None:
                 mutated.write_text(mutated.read_text().replace(before, after))
                 problems = validator.check_js_adapters(Path(tmp))
                 self.assertTrue(any(filename in problem for problem in problems), problems)
+
+
+class ScriptTestsDiscoveryContractTest(unittest.TestCase):
+    def test_all_script_tests_are_discoverable(self) -> None:
+        scripts_tests_dir = validator.ROOT / "scripts" / "tests"
+        test_files = sorted(f.stem for f in scripts_tests_dir.glob("test_*.py"))
+        self.assertTrue(
+            len(test_files) >= 13, f"Expected at least 13 test modules, found {len(test_files)}"
+        )
+
+        loader = unittest.defaultTestLoader
+        suite = loader.discover(str(scripts_tests_dir), top_level_dir=str(scripts_tests_dir))
+
+        loaded_modules: set[str] = set()
+
+        def _walk_suite(s: Any) -> None:
+            for item in s:
+                if isinstance(item, unittest.TestSuite):
+                    _walk_suite(item)
+                elif isinstance(item, unittest.TestCase):
+                    loaded_modules.add(item.__class__.__module__)
+
+        _walk_suite(suite)
+        for mod in test_files:
+            self.assertIn(mod, loaded_modules, f"Module {mod} was not discovered by unittest")
+
+    def test_workflows_and_agents_use_discovery_not_hand_kept_lists(self) -> None:
+        quality_gate_path = validator.ROOT / ".github" / "workflows" / "quality-gate.yml"
+        quality_gate_text = quality_gate_path.read_text(encoding="utf-8")
+        agents_path = validator.ROOT / "AGENTS.md"
+        agents_text = agents_path.read_text(encoding="utf-8")
+
+        # Must not contain hand-enumerated scripts.tests module lists
+        self.assertNotIn("scripts.tests.test_bump_version", quality_gate_text)
+        self.assertNotIn("scripts.tests.test_bump_version", agents_text)
+
+        # Must use discover pattern for scripts/tests
+        self.assertIn("unittest discover -s scripts/tests -t scripts/tests", quality_gate_text)
+        self.assertIn("unittest discover -s scripts/tests -t scripts/tests", agents_text)