Remove pointless catalog-literal tests; document the anti-pattern

Two tests added while triaging #244/#226 just re-asserted a translation
string against the catalog entry that had been written moments earlier
(dryer_cycle_table_03's '51'/'53'/'4e', dishwasher_cycle's '83'/'86').
Neither exercises any code path -- they pass by construction and only
break when someone later edits the label text for wording, not when the
actual code/value mapping regresses. tests/test_translations.py already
holds the invariants that matter for catalog data.

Documents the anti-pattern in the adding-device-support skill so future
translation-only fixes don't reach for this pattern again.

.claude/skills/adding-device-support/SKILL.md

@@ -399,6 +399,24 @@ no `[%key:...%]` resolution (that's Core build tooling). Every other language
 must mirror `en.json` key for key — also enforced by
 `tests/test_translations.py`.
 
+**Don't write a test that just re-asserts a translation string.** Adding
+labels is a data change, not a logic change, and `tests/test_translations.py`
+already holds the invariants that matter for data (every descriptor has a
+catalog entry, every language mirrors English key-for-key, no unresolved
+`[%key:...%]`). A test that loads the catalog and asserts
+`catalog["select"]["foo"]["state"]["16"] == "Cotton"` right after you just
+wrote that exact line into `en.json` doesn't exercise any code path — it
+re-states the JSON file in Python, passes by construction, and only ever
+fails when someone *correctly* edits the label later (a wording fix, a
+translator's improvement). It's not a regression test, because there's no
+`select.py`/`adapter.py` logic between "the JSON says X" and "the test reads
+X" for it to catch drift in. If a code/label mapping is worth locking in,
+test it through the code that actually consumes it instead — a write
+contract (`desc.write_fn(...)` returns the right raw code), a read contract
+(`flatten()` produces the right raw value from a fixture rep), or a routing
+decision — never a bare literal-string comparison against the catalog you
+just edited.
+
 ## 8. Coverage discipline: bound or ignored
 
 Every href in the dump must resolve, or the repair fires. If a resource isn't

tests/test_dishwasher_capabilities.py

@@ -33,17 +33,6 @@ class TestCycleOptions:
         assert body == {"x.com.samsung.da.options": ["Course_90"]}
 
 
-def test_normal_and_express_60_labels_are_not_transposed():
-    """Issue #226: '83'/'86' were swapped in the shipped catalog -- '86' is
-    Express 60, '83' is Normal (see dishwasher.py's CYCLE_OPTIONS comment for
-    how this was confirmed)."""
-    from custom_components.localthings.catalog import _ENTITY_CATALOG
-
-    state = _ENTITY_CATALOG["select"]["dishwasher_cycle"]["state"]
-    assert state["83"] == "Normal"
-    assert state["86"] == "Express 60"
-
-
 class TestDishwasherOptions:
     def test_storm_wash_read_and_write(self):
         desc = next(

tests/test_dryer_capabilities.py

@@ -94,15 +94,3 @@ def test_st_dryercourse_is_ignored():
     ignored_hrefs = {c.href for c in ignored.IGNORED}
     assert "/st/dryercourse/vs/0" in ignored_hrefs
     assert "/st/washercourse/vs/0" in ignored_hrefs
-
-
-def test_table_03_labels_confirmed_on_dv90dg6845lhu5():
-    """Issue #244: codes 51/53/4e were confirmed by selecting each program
-    directly on the physical DV90DG6845LHU5 and reading back the resulting
-    raw course code -- same verification method as issue #80's '52'/'54'/'60'
-    on the washer's own course table."""
-    from custom_components.localthings.catalog import translated_states
-
-    known = translated_states("select", "dryer_cycle_table_03")
-    for code in ("51", "53", "4e", "17", "21", "4c"):
-        assert code in known, code