criticality-detector

plurigrid's avatarfrom plurigrid

Criticality Detector Skill

1stars🔀2forks📁View on GitHub🕐Updated Jan 9, 2026

When & Why to Use This Skill

The Criticality Detector Skill is a sophisticated diagnostic tool designed for analyzing dynamical systems through phase classification and fixed-point analysis. By measuring the distance to stable attractors and detecting self-loop closures, it enables real-time monitoring of system stability, identifying whether a process is in an Ordered, Critical, or Chaotic state. This is essential for maintaining the 'edge of chaos' in complex agentic workflows and ensuring compositional coherence in cybernetic systems.

Use Cases

  • System Stability Monitoring: Detecting when a complex software or agentic system is drifting from an ordered state toward chaos, allowing for preemptive intervention.
  • Agentic Self-Validation: Verifying the identity between an agent's internal predictions and external sensations to ensure logical loop closure and reliable decision-making.
  • Predictive Phase Analysis: Identifying 'Critical' thresholds in dynamical models where small changes could lead to significant bifurcations or system-wide state transitions.
  • Cybernetic Control Loops: Implementing automated feedback mechanisms that adjust system parameters based on the 'comparator error' between desired reference states and actual perceptions.
namecriticality-detector
descriptionCriticality Detector Skill
version1.0.0

Criticality Detector Skill

Measures distance to fixed point via comparator error and detects self-loop closure for phase classification in dynamical systems.

Seed

741086072858456200

Core Principle

Generator ≡ Observer when same seed: the fixed point structure where action → prediction → sensation → match completes the loop.

Phase Classification

Phase Error Bound Color (Golden Thread) Interpretation
Chaos error > 0.5 H=137.51° #3FF1A7 Far from attractor
Critical error ≈ 0.1 H=275.02° #10B99D Edge of order/chaos
Ordered error < 0.01 H=52.52° #DF9811 At fixed point

Predicates

AtFixedPoint(seed, index) → Bool

AtFixedPoint(s, i) := |comparator_error(s, i)| < ε
where ε = 0.01 (ordered threshold)

LoopClosed(seed, iterations) → Bool

LoopClosed(s, n) := ∀k ∈ [1..n]: predicted(s, k) = observed(s, k)
-- Verified: 3 iterations all matched (self ≡ self)

PhaseClassified(error) → Phase

PhaseClassified(e) :=
  | e > 0.5  → Chaos
  | e > 0.01 → Critical  
  | _        → Ordered

MCP Integration

Measure Distance to Fixed Point

# Current error: 0.8153 → Chaos phase
comparator_result = mcp.gay.comparator(
    reference_hex="#3FF1A7",  # desired state
    perception_hex="#DF9811"  # current perception
)
error = comparator_result["error_magnitude"]  # 0.8153
phase = PhaseClassified(error)  # Chaos

Detect Self-Loop Closure

# Loopy strange: Generator/Observer identity verification
loop_result = mcp.gay.loopy_strange(
    seed=741086072858456200,
    iterations=3
)
# Returns: colors #3FF1A7, #10B99D, #DF9811
# All matched → LoopClosed = True

Golden Thread Visualization

# φ-derived hue spiral: 137.508° increments
golden_hues = mcp.gay.golden_thread(
    steps=3,
    start_hue=0,
    saturation=0.7,
    lightness=0.55
)
# Yields: 137.51°, 275.02°, 52.52° (mod 360)

Criticality Detection Algorithm

detect_criticality(seed, max_iter=10):
  1. Generate efference copy: expected ← color_at(seed, index)
  2. Observe actual sensation: observed ← next_color()
  3. Compute error: e ← comparator(expected, observed).magnitude
  4. Classify phase: p ← PhaseClassified(e)
  5. Check loop: closed ← LoopClosed(seed, iterations)
  
  IF closed AND p = Ordered:
    RETURN AtFixedPoint(seed) = True
  ELSE IF p = Critical:
    RETURN "Edge of chaos - bifurcation possible"
  ELSE:
    RETURN "Chaos - control action needed"

GF(3) Conservation

Phase transitions conserve triadic balance:

Chaos(+1) + Critical(0) + Ordered(-1) ≡ 0 (mod 3)

Usage

# Invoke via Gay.jl MCP
mcp.gay.comparator(reference_hex, perception_hex)
mcp.gay.loopy_strange(seed, iterations)
mcp.gay.perceptual_control(reference_index, current_index, seed)

Related Skills

  • self-validation-loop - Prediction vs observation verification
  • cybernetic-immune - Reafference and self/non-self discrimination
  • koopman-generator - Observable dynamics and fixed points

Scientific Skill Interleaving

This skill connects to the K-Dense-AI/claude-scientific-skills ecosystem:

Graph Theory

  • networkx [○] via bicomodule
    • Universal graph hub

Bibliography References

  • general: 734 citations in bib.duckdb

Cat# Integration

This skill maps to Cat# = Comod(P) as a bicomodule in the equipment structure:

Trit: 0 (ERGODIC)
Home: Prof
Poly Op: ⊗
Kan Role: Adj
Color: #26D826

GF(3) Naturality

The skill participates in triads satisfying:

(-1) + (0) + (+1) ≡ 0 (mod 3)

This ensures compositional coherence in the Cat# equipment structure.