bench(learn): exp-001 and exp-002 registered in the memory store, units pending (decision-engine task 7.2)

Registration rows at pin fbc3220 in s3://sast-memory; in-repo records under
benchmarks/experiments/<id>/registration.json. exp-002's units could not be
frozen: every example row in the store carries null plane-only instruments
and learn.examples.load_examples raises on them (unchanged code since
a90e570) -- a store-side change to diagnose before any paid run.

Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_01EmM1j46qgKMh3zqD1QMwEo

.kiro/specs/learned-decision-engine/tasks.md

@@ -133,7 +133,10 @@ a model or embedding call.
     annotation checks / the `experiment-outcome` task are not written) and `exp-002-retrieval-ensemble` (arm A today's
     k = 5 temperature sampling, arm B the retrieval ensemble; primary canary agreement (target 0.90) and within-pair
     AUC, secondary usd per candidate, AUC, ADVISORY recall/precision; families injection and access_control on the
-    slice-2 candidates; ceiling $4: A $1, B $3).
+    slice-2 candidates; ceiling $4: A $1, B $3). Both registered in `s3://sast-memory` at fbc3220
+    (`benchmarks/experiments/<id>/registration.json`). exp-002's units are `pending` too: on 2026-10-02 every example
+    row in the store carried null plane-only instruments (`agree`, `model_sinks`, `verify`) and `load_examples` raised
+    on them, so `ousast learn experiment units` could not freeze the candidates.
   - [ ] 7.3 exp-001 run and recorded (~$7.7 expected, ceiling $15) -- an inconclusive result is pre-registered as
     likely; needs the frozen units file first.
   - [ ] 7.4 exp-002 run and recorded **(budget go-ahead: ~$2 expected, ceiling $4)**; further program variants

benchmarks/experiments/exp-001-known-callers/registration.json

@@ -0,0 +1,11 @@
+{
+  "experiment": "exp-001-known-callers",
+  "registered": "2026-10-02",
+  "store": "s3://sast-memory",
+  "manifest": "plane/experiments/exp-001-known-callers.yaml",
+  "manifest_digest": "76b54aa5ae908a4cd7797f3b4654acf03b90e5ad5d068216036e722104313929",
+  "pin": "fbc3220bbe1c8a4f8afa0a724fd41a51b390dd22",
+  "units_status": "pending",
+  "units_note": "model-free candidate generation on the 41 development repositories is not built; the plane's experiment-outcome task and Run annotation checks are not written",
+  "status": "registered, not run"
+}

benchmarks/experiments/exp-002-retrieval-ensemble/registration.json

@@ -0,0 +1,11 @@
+{
+  "experiment": "exp-002-retrieval-ensemble",
+  "registered": "2026-10-02",
+  "store": "s3://sast-memory",
+  "manifest": "plane/experiments/exp-002-retrieval-ensemble.yaml",
+  "manifest_digest": "f5004fb8618b1f7cc7dcf298d42b4ac1b6c0c1aa72a97299c77f937bd6175148",
+  "pin": "fbc3220bbe1c8a4f8afa0a724fd41a51b390dd22",
+  "units_status": "pending",
+  "units_note": "`ousast learn experiment units` could not freeze the slice-2 candidates on 2026-10-02: all 904 example rows in the store carry null for the plane-only instruments (agree, model_sinks, verify) and learn.examples.load_examples (unchanged since a90e570) raises on them; freeze, commit and register again once the rows read",
+  "status": "registered, not run"
+}