fix(learning-harness): the round-3 review findings, remediated RED-first

Eight of nine group-3 tasks were rejected. Each defect was hiding a number,
and the numbers are why the loop was not learning:

- The fixed side was judged against the *vulnerable* side's function
  ranges. 191 of 227 pairs with a resolvable labeled span (84%) have
  different spans on the two sides, so a leak was recorded as silence,
  always in the direction that flatters Youden.
- Round zero measured every family with the generalist. `family:unknown`
  earns no credit against any real family, so every family but `unknown`
  baselined at a structural zero and every noise floor with it.
- The cross-family sweep counted absolute failures instead of negative
  flips. On a corpus where most pairs fail at round zero that rejects
  every round for damage it did not do. Unscorable rows voted too.
- K runs was declared and never threaded; `pairs --k-runs` was parsed and
  discarded, and its only test grepped `--help`.
- The meta-agent proposer was a stub, so `learning round` could never
  propose anything. It now composes MetaAgent.evolve against a scratch
  copy of the family directory, and `--proposals FILE` runs a recorded
  sequence offline.
- `refused_holdout` was a vocabulary word no code path could produce, and
  the covering test accepted either outcome.
- Cost was metered once, after everything had been paid for; a stage that
  raised left no record at all.
- The corrected lever number was published beside the earlier one rather
  than in place of it, and the test named for that property never checked.
- Refusals were `continue` statements: `Refusal.degradation()` existed and
  nothing called it. A train-split identical twin was reported as
  `split:train`, which is untrue.

Test repairs that each changed a number: two split tests measured an
`x = 1` fixture that teaches nothing whatever the rule says; two CLI
scoreboard tests asserted against all-zero output produced by a stub that
answers twice and puts its finding outside the labeled function.

Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01T7XD7x5KonMdmGRkp63EQG

tests/test_obligation_export.py

@@ -128,8 +128,12 @@ def test_the_train_split_of_vibe_py_teaches_no_obligation_shape_today(tmp_path:
     cases = select_vendored(select_slice(load_pair_catalog(), "vibe-py"))
     report = export_mechanisms(cases, store)
     assert obligation_mechanisms(store.load()) == ()
-    refused = {pair for pair, reason in report.skipped if reason.startswith("split:")}
-    assert {"vampi-books-get-by-title", "vampi-users-update-password", "threatbyte-api-v1-get"} <= refused
+    refused = {pair: reason for pair, reason in report.skipped}
+    assert refused["vampi-users-update-password"] == "split:holdout"
+    assert refused["threatbyte-api-v1-get"] == "split:holdout"
+    # Its own reason, not its split: this row is a byte-identical twin that happens to sit on the holdout side,
+    # and reporting `split:holdout` would hide a corpus fact behind a split fact (Req 5.2).
+    assert refused["vampi-books-get-by-title"] == "unscorable:identical_twin"
     flow_rows = [row for row in store.load() if row.shape is not None and row.shape.get("family", "sink") == "sink"]
     assert report.records == len(flow_rows) and len(flow_rows) >= 1  # the flow family still teaches from the train split