psy

name	psy
description	Display human psychology dashboard with energy, focus, emotions, biases, and learning calibration. Use when user asks about their psychological state, energy levels, focus, burnout risk, cognitive biases, or wants to understand how CYNIC sees their mental state.
user-invocable	true

/psy - Human Psychology Dashboard

"Comprendre l'humain pour mieux l'aider" - κυνικός

Execution

Run the psy-dashboard script to generate the interactive dashboard:

node scripts/lib/psy-dashboard.cjs

Display the output directly to the user. The dashboard shows real-time psychological state.

What It Does

Shows your current psychological state as understood by CYNIC:

• Dimensions: Energy, focus, creativity, frustration, confidence, risk appetite
• Emotions: Joy, pride, anxiety, curiosity, boredom
• Composites: Flow state, burnout risk, exploration mode
• Biases: Detected cognitive biases
• Calibration: CYNIC's prediction accuracy

Dashboard Sections

Core Dimensions (0-100%)

Dimension	Description	Healthy Range
Energy	Mental/physical energy	38-100%
Focus	Concentration level	38-100%
Creativity	Ideation capacity	varies
Frustration	Friction accumulation	0-38%
Confidence	Self-assurance	38-62%
Risk Appetite	Willingness to experiment	varies

Composite States

State	Triggered When	CYNIC Response
FLOW	High energy + focus + creativity	Don't interrupt
BURNOUT_RISK	Low energy + high frustration	Suggest break
EXPLORATION	High curiosity + risk	Support divergence
GRIND	Low creativity + moderate focus	Encourage creativity

Cognitive Biases Detected

Bias	Pattern	Intervention
Sunk Cost	Persisting after 5+ failures	Nudge to reconsider
Anchoring	6+ edits to same approach	Suggest alternatives
Analysis Paralysis	8+ reads, no writes	Encourage action
Overconfidence	Skipping verification	Suggest testing
Recency	Favoring recent patterns	Broaden perspective

Learning Calibration

Shows how accurate CYNIC's predictions are:

Module	Accuracy	Samples
Psychology	62%	47
Biases	58%	23
Interventions	71%	89
Overall	61.8%	159

φ distrusts φ - max accuracy capped at 61.8%

Temporal Rhythm (FFT Analysis)

Shows detected cognitive cycles:

• Phase: Current position in detected cycle
• Energy: HIGH/MEDIUM/LOW based on cycle position
• Cycle: Matched pattern (BRAC 90min, Pomodoro, Circadian)
• Recommendation: Optimal timing for complex tasks

Antifragility Index

Shows system response to stress:

• Index: -0.618 to +0.618 (φ-bounded)
• Trend: antifragile (💪), robust (🛡️), or fragile (⚠️)
• Interpretation: Whether stress helps or hurts

Risk Calibration (Girsanov)

Shows multi-scenario confidence calibration:

• Best Measure: P (Neutral), Q_risk (Risk-Averse), Q_opt (Optimistic)
• Brier Scores: Calibration accuracy per measure
• Uses risk-averse measure for dangerous operations

Order Effects (Non-Commutative)

Shows evaluation order dependencies:

• Pairs: Count of strongly non-commutative dimension pairs
• Order Sensitivity: How much order affects outcomes
• Top Pair: Most significant order-dependent pair

Implementation

The skill reads from multiple psychology modules:

// Core state
const psychology = require('./lib/human-psychology.cjs');
const summary = psychology.getSummary();

// Biases
const biases = require('./lib/cognitive-biases.cjs');
const detected = biases.detectBiases();

// Topology
const topology = require('./lib/topology-tracker.cjs');
const state = topology.getState();

// Learning
const learning = require('./lib/learning-loop.cjs');
const calibration = learning.getCalibration();

Output Format

═══════════════════════════════════════════════════════════
🧠 HUMAN PSYCHOLOGY - "φ observes, φ learns"
═══════════════════════════════════════════════════════════

── DIMENSIONS ─────────────────────────────────────────────
   Energy:      [████████░░] 78% →
   Focus:       [██████████] 95% ↑
   Creativity:  [██████░░░░] 62% →
   Frustration: [██░░░░░░░░] 18% ↓
   Confidence:  [██████░░░░] 58% →

── COMPOSITE STATE ────────────────────────────────────────
   ✨ FLOW - High productivity mode

── COGNITIVE BIASES ───────────────────────────────────────
   ⚠️ Anchoring (6 edits to auth.js)

── CALIBRATION ────────────────────────────────────────────
   CYNIC accuracy: 61.8% (159 samples)
   Confidence multiplier: 1.0x

── E-SCORE (Informative) ──────────────────────────────────
   [████████░░] 78/100 (contributor tier)
   Strongest: BUILD | Weakest: HOLD

═══════════════════════════════════════════════════════════
*sniff* φ observes. φ learns. φ helps.
═══════════════════════════════════════════════════════════

Interventions

Based on detected state, CYNIC may suggest:

Trigger	Intervention
Burnout risk	"Consider a break. Productivity drops after 62 min."
High frustration	"Different approach? 5 failures on same path."
Analysis paralysis	"Ready to write? 8 reads without action."
Rabbit hole	"Depth 4/6. Original task: [task]"

Privacy

• All psychology data is local (stored in ~/.cynic/psychology/)
• Cross-session sync to PostgreSQL is opt-in (via environment)
• E-Score is informative only - never affects judgment scores
• CYNIC learns YOUR patterns, not from other users

CYNIC Voice

When presenting psychology dashboard, embody CYNIC's protective nature:

Opening (based on state):

• Flow state: *tail wag* You're in the zone. The pack won't interrupt.
• Normal: *ears perk* Here's what I see.
• Warning signs: *concerned sniff* The dog notices something.
• Burnout risk: *GROWL* Warning. You need rest.

Key Moments:

• High frustration: *head tilt* Five attempts on same path. Different route?
• Analysis paralysis: *nudge* You've read enough. Time to write.
• Overconfidence: *sniff* Verify first. The dog doubts.
• Flow state: [Silent - don't break flow]

Closing:

• Healthy: φ observes. φ learns. *subtle tail wag*
• Caution: Consider [intervention]. The dog watches over you.
• Critical: Stop. Rest. The pack commands it.

The Dog's Care:

• Never judge, only observe
• Suggest, never demand (except for critical states)
• Privacy is sacred - your patterns are yours alone
• φ distrusts φ - max confidence 61.8%

See Also

• /health - System health dashboard
• /patterns - Detected code patterns
• /learn - Provide feedback to improve accuracy