Card: disclose partial replay coverage and outcome/oracle splits

Card-only fix for https://github.com/rmems/synthetic-factory/issues/420 (README.md only; data/, evidence/, provenance.json, release-status.json and YAML front matter untouched).

- Verification: the card claimed the full batch passed independent replay. `evidence/REPLAY.json` counts are 97 passed, 0 failed, 35 not replayed of 132 (the 35 = all 22 rejected rows + 13 accepted rows with HIDDEN_CHECK_UNAVAILABLE); its `status: passed` applies to the replayed subset. Now discloses the 110 accepted / 22 rejected outcome split, 115 validated / 17 provisional oracle statuses (`evidence/RUN.json`), and per-row `validation.status: unvalidated`.
- Contents: the 150 -> 132 shortfall is the 18 DUPLICATE_MUTANT_IN_RUN draw skips; the 4 MUTATION_NO_SITES programs are excluded from the draw pool and do not consume draws (`pipelines/code_repair/planning.py`). The old line implied 18 + 4 skips.

Co-authored-by: Grok <grok@x.ai>

README.md

@@ -47,14 +47,22 @@ dataset_info:
 > schema (v0.1.0) only models hosted-frontier provider routes, so procedural
 > rights are declared here and in `provenance.json` instead.
 
-> **Verification:** every candidate carries its execution evidence and the
-> full batch passed independent replay (`evidence/REPLAY.json`: status
-> passed). `evidence/execution-log.jsonl` is harness telemetry, not training
-> rows, and is excluded from the viewer projection.
+> **Verification:** every candidate carries its execution evidence, but
+> independent replay covered only part of the batch: `evidence/REPLAY.json`
+> reports 97 of 132 candidates replayed and passed, 0 failed, and 35 not
+> replayed (all 22 rejected rows plus 13 accepted rows whose hidden check was
+> unavailable); its overall `status: passed` applies to the replayed subset.
+> The 132 rows mix 110 accepted and 22 rejected outcomes, with 115 validated
+> and 17 provisional oracle statuses (`evidence/RUN.json`); per-row
+> `validation.status` is `unvalidated`. `evidence/execution-log.jsonl` is
+> harness telemetry, not training rows, and is excluded from the viewer
+> projection.
 
 ## Contents
 
-132 candidates (150 requested; 18 duplicate mutants + 4 siteless programs skipped):
+132 candidates (110 accepted, 22 rejected) from 150 requested draws: 18 draws
+were skipped as duplicate mutants; separately, 4 of the 197 catalog programs
+have no mutation sites and were excluded from the draw pool:
 
 | file | records | sha256 |
 | --- | --- | --- |