feat(profiler/behave_shell): emit motor.paste_burst_rate

BEHAVE-EXTRACTOR.md Phase A Step 3. Same paste-event ratio as
motor.input_modality but coarser-bucketed: this is the *habit*
signal (does the operator reach for paste at all?), where
input_modality is the dominant-channel signal.

* _features/motor.py:paste_burst_rate(ctx) emits one Observation
  per session in {none, occasional, habitual} with confidence
  0.70 / 0.70 / 0.80.
* Thresholds: PASTE_RATE_OCCASIONAL_MIN=0.10,
  PASTE_RATE_HABITUAL_MIN=0.50.

Splits YOU-sim from LW/CLAUDE-FF/CLAUDE-CL — LLM-driven sessions
paste habitually, real humans rarely paste.

Tests: pure-typed → none; 1-paste-in-10 → occasional;
paste-majority → habitual; output-only → no observation; habitual
confidence > occasional confidence.

decnet/profiler/behave_shell/_features/__init__.py

@@ -11,10 +11,14 @@ from typing import Callable, Iterable
 from decnet_behave_core.spec.envelope import Observation
 
 from decnet.profiler.behave_shell._ctx import SessionContext
-from decnet.profiler.behave_shell._features.motor import input_modality
+from decnet.profiler.behave_shell._features.motor import (
+    input_modality,
+    paste_burst_rate,
+)
 
 FeatureFn = Callable[[SessionContext], Iterable[Observation]]
 
 FEATURES: tuple[FeatureFn, ...] = (
     input_modality,
+    paste_burst_rate,
 )

decnet/profiler/behave_shell/_features/motor.py

@@ -14,6 +14,8 @@ from decnet.profiler.behave_shell._features._emit import make_observation
 from decnet.profiler.behave_shell._thresholds import (
     MODALITY_PASTED_MIN,
     MODALITY_TYPED_MAX,
+    PASTE_RATE_HABITUAL_MIN,
+    PASTE_RATE_OCCASIONAL_MIN,
 )
 
 
@@ -43,3 +45,34 @@ def input_modality(ctx: SessionContext) -> Iterator[Observation]:
         value=modality,
         confidence=confidence,
     )
+
+
+def paste_burst_rate(ctx: SessionContext) -> Iterator[Observation]:
+    """Emit ``motor.paste_burst_rate`` ∈ {none, occasional, habitual}.
+
+    Same paste-event ratio as ``input_modality`` but coarser-bucketed:
+    this primitive is the *habit* signal (does the operator reach for
+    paste at all?), where input_modality is the dominant-channel
+    signal (is the session paste-driven overall?). Splits YOU-sim from
+    LW/CLAUDE-FF/CLAUDE-CL — LLM-driven sessions paste habitually,
+    real humans don't.
+    """
+    n = len(ctx.input_events)
+    if n == 0:
+        return
+    ratio = ctx.paste_event_count / n
+    if ratio >= PASTE_RATE_HABITUAL_MIN:
+        level = "habitual"
+        confidence = 0.80
+    elif ratio >= PASTE_RATE_OCCASIONAL_MIN:
+        level = "occasional"
+        confidence = 0.70
+    else:
+        level = "none"
+        confidence = 0.70
+    yield make_observation(
+        ctx,
+        primitive="motor.paste_burst_rate",
+        value=level,
+        confidence=confidence,
+    )