feat(harness): intentional skips + custom-html-tiny functional test; 4-rung ladder (#6)

Declare intentional skips + custom-html-tiny functional test; 4-rung level ladder

- recipe_meta.EXPECTED_NA = {rung: reason} lists intentionally-skipped rungs; any
  essential rung skipped and not listed is unintentional. Skips still cap the level
  (never inflate). results.json: skips:{intentional,unintentional} + level_cap_rung.
- Level ladder = the four essential rungs (install, upgrade, backup/restore,
  functional; top = L4). integration & recipe-local are optional, not leveled
  (SSO still enforced for the run verdict, unchanged).
- Card shows skipped rungs as INTENTIONAL SKIP (green, reason below) / UNINTENTIONAL
  SKIP (amber); level badge gains an expected/gap? third segment.
- custom-html-tiny: functional serve test (exact-byte round-trip + 404); declares
  backup_restore intentionally skipped (stateless static server).

Independently verified by the adversary: 138 unit tests pass cold; live full-stage
run on custom-html-tiny green (upgrade tier ran; level 2; correct skips/badge);
clean teardown.

tests/unit/test_results.py

@@ -105,83 +105,31 @@ def _results(**kw):
     return base
 
 
-def test_derive_rungs_full_stateful_sso():
-    rungs = R.derive_rungs(
-        _results(),
-        backup_capable=True,
-        declared=["keycloak"],
-        deps_ready=True,
-        sso_unverified=False,
-        has_custom=True,
-        has_repo_local=False,
-        repo_local_passed=False,
-    )
+def test_derive_rungs_full_climb_four_essential():
+    rungs = R.derive_rungs(_results(), backup_capable=True, has_custom=True)
+    # only the four essential rungs — integration/recipe-local are optional, not produced here.
     assert rungs == {
         "install": "pass",
         "upgrade": "pass",
         "backup_restore": "pass",
         "functional": "pass",
-        "integration": "pass",
-        "recipe_local": "na",
     }
 
 
-def test_derive_rungs_no_sso_surface_is_integration_na():
-    rungs = R.derive_rungs(
-        _results(),
-        backup_capable=True,
-        declared=[],
-        deps_ready=True,
-        sso_unverified=False,
-        has_custom=True,
-        has_repo_local=False,
-        repo_local_passed=False,
-    )
-    assert rungs["integration"] == "na"
-    assert rungs["functional"] == "pass"
-
-
-def test_derive_rungs_stateless_backup_na():
+def test_derive_rungs_stateless_backup_and_functional_na():
     rungs = R.derive_rungs(
         _results(backup="skip", restore="skip", custom="skip"),
         backup_capable=False,
-        declared=[],
-        deps_ready=True,
-        sso_unverified=False,
         has_custom=False,
-        has_repo_local=False,
-        repo_local_passed=False,
     )
     assert rungs["backup_restore"] == "na"
     assert rungs["functional"] == "na"
+    assert "integration" not in rungs and "recipe_local" not in rungs
 
 
-def test_derive_rungs_sso_unverified_is_integration_fail():
-    rungs = R.derive_rungs(
-        _results(),
-        backup_capable=True,
-        declared=["keycloak"],
-        deps_ready=False,
-        sso_unverified=True,
-        has_custom=True,
-        has_repo_local=False,
-        repo_local_passed=False,
-    )
-    assert rungs["integration"] == "fail"
-
-
-def test_derive_rungs_repo_local_pass():
-    rungs = R.derive_rungs(
-        _results(),
-        backup_capable=True,
-        declared=[],
-        deps_ready=True,
-        sso_unverified=False,
-        has_custom=True,
-        has_repo_local=True,
-        repo_local_passed=True,
-    )
-    assert rungs["recipe_local"] == "pass"
+def test_derive_rungs_functional_fail():
+    rungs = R.derive_rungs(_results(custom="fail"), backup_capable=True, has_custom=True)
+    assert rungs["functional"] == "fail"
 
 
 # ---- build_results: end-to-end incl level + flags ----
@@ -212,16 +160,13 @@ def test_build_results_level_and_flags(tmp_path):
         records=recs,
         results=_results(),
         backup_capable=True,
-        declared=[],
-        deps_ready=True,
-        sso_unverified=False,
         clean_teardown=True,
         no_secret_leak=True,
         finished_ts=1234.0,
     )
-    # stateful, functional pass, no SSO surface, no repo-local → caps at L4
+    # all four essential rungs pass → full climb to L4 (the top), no cap
     assert data["level"] == 4
-    assert "L5" in data["level_cap_reason"]
+    assert data["level_cap_reason"] == ""
     assert data["recipe"] == "hedgedoc"
     assert data["ref"] == "deadbeefcafe"
     assert data["flags"] == {"clean_teardown": True, "no_secret_leak": True}
@@ -246,9 +191,6 @@ def test_build_results_capped_at_L1_on_upgrade_fail(tmp_path):
         records=recs,
         results=_results(upgrade="fail"),
         backup_capable=True,
-        declared=[],
-        deps_ready=True,
-        sso_unverified=False,
         clean_teardown=True,
         no_secret_leak=True,
         finished_ts=0.0,
@@ -257,6 +199,85 @@ def test_build_results_capped_at_L1_on_upgrade_fail(tmp_path):
     assert "L2" in data["level_cap_reason"]
 
 
+# ---- skips: intentional (declared) vs unintentional (everything else skipped) ----
+
+
+def _rungs(**kw):
+    base = {
+        "install": "pass",
+        "upgrade": "pass",
+        "backup_restore": "pass",
+        "functional": "pass",
+    }
+    base.update(kw)
+    return base
+
+
+def test_skips_intentional_vs_unintentional():
+    rungs = _rungs(backup_restore="na", functional="na")
+    sk = R.skips(rungs, {"backup_restore": "stateless static server"})
+    # backup_restore is declared (intentional, with reason); functional skipped but not declared.
+    assert sk["intentional"] == {"backup_restore": "stateless static server"}
+    assert sk["unintentional"] == ["functional"]
+
+
+def test_skips_none_declared_all_unintentional():
+    rungs = _rungs(backup_restore="na")
+    sk = R.skips(rungs, None)
+    assert sk["intentional"] == {}
+    assert sk["unintentional"] == ["backup_restore"]
+
+
+def test_skips_declaration_only_counts_when_actually_skipped():
+    # backup_restore actually ran (pass) → not a skip, so a declaration for it is simply inert.
+    rungs = _rungs(backup_restore="pass")
+    sk = R.skips(rungs, {"backup_restore": "reason"})
+    assert "backup_restore" not in sk["intentional"]
+    assert "backup_restore" not in sk["unintentional"]
+
+
+def test_build_results_threads_expected_na(tmp_path):
+    # Mirrors custom-html-tiny post-change: install + a passing functional (custom) test, but no
+    # backup surface (backup_restore declared intentionally skipped).
+    recs = [
+        {
+            "tier": "install",
+            "source": "generic",
+            "file": "g/test_install.py",
+            "rc": 0,
+            "junit": _write(tmp_path, "i.xml", JUNIT_PASS),
+        },
+        {
+            "tier": "custom",
+            "source": "cc-ci",
+            "file": "c/test_serves_content.py",
+            "rc": 0,
+            "junit": _write(tmp_path, "c.xml", JUNIT_PASS),
+        },
+    ]
+    data = R.build_results(
+        recipe="custom-html-tiny",
+        version="1.1.0",
+        pr="0",
+        ref=None,
+        records=recs,
+        results=_results(backup="skip", restore="skip"),  # custom=pass (default) → functional pass
+        backup_capable=False,  # no backupbot label → backup_restore skipped (N/A)
+        clean_teardown=True,
+        no_secret_leak=True,
+        finished_ts=0.0,
+        expected_na={"backup_restore": "stateless static file server"},
+    )
+    # backup_restore skip still caps at L2 (never inflates) — even though functional passes above it,
+    # the skip caps the climb — but it's the declared (intentional) rung that capped.
+    assert data["level"] == 2
+    assert "L3" in data["level_cap_reason"]
+    assert data["level_cap_rung"] == "backup_restore"
+    assert data["rungs"]["functional"] == "pass"
+    assert data["skips"]["intentional"]["backup_restore"] == "stateless static file server"
+    assert data["skips"]["unintentional"] == []  # backup_restore declared; functional passed → clean
+
+
 def test_write_results_roundtrip(tmp_path):
     data = {"run_id": "42", "level": 3, "stages": []}
     path = R.write_results(data, runs_dir_override=str(tmp_path))