docs(v1.8.1): gameplay_metrics validated end-to-end + harness-behaviour guidance

.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "ooda-loop",
   "displayName": "OODA-loop",
-  "version": "1.8.0",
+  "version": "1.8.1",
   "description": "An autonomous operations layer for your live side project. It watches, re-orients from which PRs you merge and reject, and opens small revertible PRs — bounded by a HALT file, protected paths, and a hard cost cap. Built on Boyd's OODA loop. You stay in command.",
   "author": {
     "name": "Taeil Ma",

CHANGELOG.md

@@ -8,6 +8,30 @@ independently. Bump there signals migration work for downstream projects.
 
 ---
 
+## [v1.8.1] — 2026-06-19
+
+### Validated + guidance — the gameplay_metrics path works end-to-end
+
+The f1 probe exercised v1.8.0's per-dimension capture: it authored a
+`gameplay_metrics` harness (drives the real pure physics headlessly) for the two
+**frozen** experiential axes (`driving_feel`, `fun_challenge` = 45% of the rubric
+a screenshot can't judge).
+
+- **Honest measurement first dropped artifact_quality 0.533 → 0.490** — the
+  screenshot had been *over*-scoring feel/fun (0.51/0.38 → measured 0.41/0.29).
+  Confirms the v1.8.0 thesis: measurement was the bottleneck *and* inflating.
+- Two leaps the unlock enabled: `fun_challenge` 0.29 → **0.81** (distinct AI
+  racing lines + tamed DRS slingshot) and `driving_feel` 0.41 → **0.78** (steering
+  inertia + power oversteer + weight transfer). **artifact_quality crossed the bar
+  for the first time (0.687 ≥ 0.65) → an HONEST grade A** (the loop's original A
+  was a lie; this one is earned).
+- **Guidance (config doc):** a `gameplay_metrics` harness must MEASURE BEHAVIOUR
+  (drive the artifact, read the numbers), never assert an implementation fact — the
+  probe's first harness hardcoded a feature flag and couldn't credit the fix,
+  which would trigger a spurious thrashing-HALT. Rewritten to measure behaviour.
+
+---
+
 ## [v1.8.0] — 2026-06-19
 
 ### Changed — drive quality to "good", not "passable" (config schema 1.4.0)

config.example.json

@@ -270,7 +270,7 @@
     "plateau_window": 4,
     "plateau_eps": 0.05,
     "locked": true,
-    "__dimensions_doc__": "v1.8.0: each dimension may override capture_method so the critic gets the evidence it actually needs. 'screenshot' axes share one capture; EXPERIENTIAL axes (feel/fun/responsiveness) a screenshot cannot judge use 'gameplay_metrics' — a HUMAN-AUTHORED harness that exercises the artifact and emits metrics JSON. The harness MUST be in safety.protected_paths AND match gameplay_metrics_hash (independence gate, same invariant as the rubric hash); else the dimension scores null (capture_failure) rather than faking a score. Without per-dimension capture, experiential axes freeze at their initial score and silently cap artifact_quality.",
+    "__dimensions_doc__": "v1.8.0: each dimension may override capture_method so the critic gets the evidence it actually needs. 'screenshot' axes share one capture; EXPERIENTIAL axes (feel/fun/responsiveness) a screenshot cannot judge use 'gameplay_metrics' — a HUMAN-AUTHORED harness that exercises the artifact and emits metrics JSON. The harness MUST be in safety.protected_paths AND match gameplay_metrics_hash (independence gate, same invariant as the rubric hash); else the dimension scores null (capture_failure) rather than faking a score. Without per-dimension capture, experiential axes freeze at their initial score and silently cap artifact_quality. v1.8.1 rule (validated by the f1 probe): the harness must MEASURE BEHAVIOUR (e.g. drive the real physics and read the resulting numbers), NOT assert an implementation fact — a hardcoded flag like {feature: false} cannot credit a real fix, so it would trigger a spurious thrashing-HALT. Drive the artifact and report what it actually does.",
     "dimensions": [],
     "__example_dimension__": {
       "name": "driving_feel",