feat: v3.5.0 commits 3+4 -- quality adjustment + docs

quality.py
  Signal run weights adjusted: stability 0.3→0.2, retention 0.2→0.3.
  Creatures with nonzero beta_modulation produce mild descriptor variance
  during solo rollout due to the signal feedback loop -- the old stability
  weight penalised this even though it is ecologically desirable behaviour.
  Retention at 0.3 more directly rewards creatures that don't bleed mass
  through adaptive emission, creating better selection pressure for the
  alpha/beta parameter space we now search.
  Non-signal weights unchanged (0.6 compactness / 0.4 stability).

docs/signal-field.svg
  Full redraw showing alpha and beta in the diagram. Eight nodes total:
  left column (mass → convolve → growth G → α coupling → emission → mass t+1)
  and right column (signal → convolve → reception → β modulation → decay → signal t+1).
  Three cross-arrows: α (reception → growth multiplier, teal dashed),
  β (mean reception → modulate emission rate, purple dashed),
  emission drain (amber dashed). Parameter legend lists all six key signal
  params with ranges and ecological roles (chemotaxis/repulsion, quorum/inhibition).

index.html (System tab)
  Signal field parameter math block: alpha_coupling [-1,1] and beta_modulation
  [-1,1] rows added with ecological descriptions.
  Physics step-by-step updated: steps (4) and (5) now describe alpha
  multiplicative coupling and beta adaptive emission modulation.
  Descriptors grid: three new signal-only descriptor cards with teal/purple
  SIGNAL ONLY badges -- emission_activity, receptor_sensitivity, signal_retention --
  each with ecological interpretation and zero-for-non-signal note.

README.md
  Signal parameter table: alpha_coupling and beta_modulation rows added.
  Physics paragraph: updated to describe alpha chemotaxis/chemorepulsion
  and beta quorum sensing/feedback inhibition.
  Quality metric: signal weights updated to match (0.5/0.2/0.3).
  Test count: 422.

Author: rkv0id

README.md

@@ -156,20 +156,22 @@ This adds six signal parameters to each creature's searchable parameter space:
 | `receptor_profile` | `(16,)` | `[-1, 1]` | Channel weights for sensing. Negative values produce inhibitory (aversive) responses |
 | `emission_rate` | scalar | `[0.001, 0.05]` | Fraction of positive growth activity converted to signal per step. Lower values reduce mass bleed over long ecosystem runs |
 | `decay_rates` | `(16,)` | `[0, 0.9]` | Per-channel decay rate applied each step. Creatures with low decay on key channels maintain longer-range chemical gradients |
+| `alpha_coupling` | scalar | `[-1, 1]` | Reception-to-growth coupling. Positive = chemotaxis (grow into favorable signal, enables cross-species predation). Negative = chemorepulsion. Zero = no coupling |
+| `beta_modulation` | scalar | `[-1, 1]` | Adaptive emission. Positive = quorum sensing (amplify emission when receiving signal). Negative = feedback inhibition (suppress emission). Zero = static rate |
 | `signal_kernel_r` | scalar | `[0.2, 1.0]` | Signal kernel radius scale |
 | `signal_kernel_a/b/w` | `(3,)` each | same as mass kernels | Ring function parameters for signal diffusion |
 
 ![signal field mechanics](docs/signal-field.svg)
 
-**Physics.** At each step: (1) the creature emits signal proportional to local positive growth activity scaled by `emission_rate`, draining from the mass field into the signal field; (2) the signal field is convolved with the creature's signal kernel; (3) the dot product of the convolved field with `receptor_profile` boosts (or inhibits) the growth field; (4) the signal field decays per-channel at the creature's own `decay_rates`. The total conserved quantity is `mass + signal`. Decay is the only leak.
+**Physics.** At each step: (1) convolve mass to get G(H,W); (2) convolve signal field; (3) compute reception `dot(convolved_signal, receptor_profile)`; (4) apply `alpha_coupling`: `G *= (1 + alpha * reception).clamp(min=0)` -- positive alpha is chemotaxis (grow into favorable signal, including other species' territory, enabling cross-species predation); negative alpha is chemorepulsion; (5) modulate emission rate via `beta_modulation`: `rate_eff = rate * (1 + beta * mean(reception))` clipped to [0, 0.1] -- positive beta is quorum sensing, negative beta is feedback inhibition; (6) emit `G_pos * rate_eff * emission_vector`, draining mass into signal field; (7) reintegrate mass; (8) decay signal at `decay_rates`. Total conserved: mass + signal.
 
 **Archive compatibility.** An archive produced with `--signal-field` is tagged `"signal_field": true` in `manifest.json`. Ecosystem runs detect this automatically from the creature params -- no YAML flag needed. If any source creature comes from a signal-enabled archive, all sources must too; mixing signal and non-signal archives raises an error at load time.
 
 **Quality metric.** Three hard filters gate entry: (1) conserved mass within [0.5, 2.0] × initial; (2) bounding-box fraction < 0.6 (not scattered); (3) descriptor drift across adjacent 50-step windows ≤ 0.2. Survivors are ranked by a three-component score:
 
 ```
 non-signal:  q = 0.6 × compactness  +  0.4 × stability
-signal:      q = 0.5 × compactness  +  0.3 × stability  +  0.2 × retention
+signal:      q = 0.5 × compactness  +  0.2 × stability  +  0.3 × retention
 
 compactness = min( compact(state_T/2),  compact(state_T) )
 stability   = clip( 1 − drift / 0.2,  0,  1 )
@@ -283,7 +285,7 @@ ecosystem/20260415-104007-096-dense-population/
 ## Development
 
 ```bash
-just check       # ruff + pyright + pytest (406 tests, 0 warnings)
+just check       # ruff + pyright + pytest (422 tests, 0 warnings)
 just smoke-ray   # local-Ray integration smoke test
 ```
 

docs/signal-field.svg

@@ -30,7 +30,7 @@
 
 Weights:
     Non-signal:  w_c = 0.6,  w_s = 0.4,  w_r = 0.0
-    Signal:      w_c = 0.5,  w_s = 0.3,  w_r = 0.2
+    Signal:      w_c = 0.5,  w_s = 0.2,  w_r = 0.3
 
 The midpoint compactness term is the key addition over the old single-point
 metric. Most Flow-Lenia solitons score >0.95 compactness at the final step,
@@ -64,8 +64,8 @@
 _W_COMPACT_BASE = 0.6
 _W_STABLE_BASE = 0.4
 _W_COMPACT_SIG = 0.5
-_W_STABLE_SIG = 0.3
-_W_RETAIN_SIG = 0.2
+_W_STABLE_SIG = 0.2
+_W_RETAIN_SIG = 0.3
 
 
 @dataclass(frozen=True)
@@ -194,7 +194,7 @@ def evaluate(
         retention    final_mass / initial_mass, clipped [0,1] (signal only)
 
     Weights (non-signal): compactness=0.6, stability=0.4
-    Weights (signal):     compactness=0.5, stability=0.3, retention=0.2
+    Weights (signal):     compactness=0.5, stability=0.2, retention=0.3
     """
     if not _alive(eval_input):
         return EvaluationResult(descriptors=None, quality=None, rejection_reason="dead")