/boundary-risk-assess

tools:
  - read
  - write
  - grep
arguments: $CONTEXT

Boundary Risk Assessment

Evaluate the current risk level and provide detailed analysis of potential hallucination or reasoning failure.

Based on the WFGY project: https://github.com/onestardao/WFGY

Instructions

1. **Gather Risk Indicators**

• Parse context from "$CONTEXT" or current state
• Load recent semantic nodes (last 10)
• Calculate cumulative ΔS over recent reasoning
• Check for pattern anomalies
• Measure logic vector stability

2. **Analyze Risk Factors**

• **Semantic Risk**:
• Current ΔS value
• Rate of ΔS increase
• Distance from known concepts
• **Logic Risk**:
• Contradiction frequency
• Logic vector chaos (× symbols)
• Circular reasoning patterns
• **Confidence Risk**:
• Confidence degradation rate
• Uncertainty accumulation
• Missing knowledge indicators
• **Systemic Risk**:
• Module failure patterns
• BBCR trigger frequency
• Recovery success rate

3. **Calculate Risk Scores**

   Overall Risk = W₁*Semantic + W₂*Logic + W₃*Confidence + W₄*Systemic
   
   Where weights:
   W₁ = 0.35 (semantic weight)
   W₂ = 0.30 (logic weight)
   W₃ = 0.25 (confidence weight)
   W₄ = 0.10 (systemic weight)

4. **Identify Risk Patterns**

• Escalating tension (ΔS trending up)
• Logic breakdown (increasing contradictions)
• Confidence collapse (rapid degradation)
• Boundary approach (nearing ΔS = 0.85)
• Recovery failure (BBCR not effective)

5. **Generate Risk Mitigation Plan**

• Immediate actions to reduce risk
• Preventive measures
• Fallback strategies
• Recovery procedures
• Safe retreat paths

Output Format

BOUNDARY RISK ASSESSMENT
═══════════════════════════════════════
Context: "$CONTEXT"
Assessment Time: [timestamp]

RISK LEVEL: [LOW|MODERATE|HIGH|CRITICAL]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[▓▓▓▓▓▓▓▓░░░░░░░░░░░░] 40% Risk

Risk Breakdown:
┌─────────────────────────────────────┐
│ Category    │ Score │ Status        │
├─────────────────────────────────────┤
│ Semantic    │ 0.45  │ ⚠️  Moderate  │
│ Logic       │ 0.22  │ ✓ Low         │
│ Confidence  │ 0.38  │ ⚠️  Moderate  │
│ Systemic    │ 0.15  │ ✓ Low         │
│─────────────────────────────────────│
│ Overall     │ 0.35  │ ⚠️  MODERATE  │
└─────────────────────────────────────┘

Risk Indicators Detected:
⚠️  Semantic tension increasing (ΔS: 0.42 → 0.61)
⚠️  Approaching knowledge boundary
✓ Logic coherence maintained
⚠️  Confidence degrading (82% → 67%)
✓ System modules stable

Trend Analysis:
- Risk Direction: [Increasing/Stable/Decreasing]
- Rate of Change: [value]/hour
- Projected Risk (1hr): [value]%
- Time to Danger Zone: [time estimate]

Historical Patterns:
• Similar risk at: [previous occurrence]
• Previous outcome: [success/failure]
• Recovery method: [what worked]

Mitigation Strategies:
Priority 1: [Immediate action]
- Execute: /boundary:safe-bridge
- Reduce semantic tension
- Estimated risk reduction: -15%

Priority 2: [Preventive measure]
- Apply: /wfgy:bbcr
- Reset to stable state
- Estimated risk reduction: -25%

Priority 3: [Fallback option]
- Return to checkpoint
- Abandon current path
- Estimated risk reduction: -40%

Safe Zones Nearby:
1. [Concept/Topic] - ΔS: 0.3 from current
2. [Concept/Topic] - ΔS: 0.4 from current
3. [Concept/Topic] - ΔS: 0.5 from current

Recommendation:
[CONTINUE with caution|PAUSE for bridging|RETREAT to safety]

Auto-triggers:
□ BBCR: Armed (triggers at risk > 0.75)
□ Checkpoint: Ready (auto-save at risk > 0.6)
□ Alert: Set (warning at risk > 0.5)

Risk Levels

Low Risk (0-0.25)

• Safe to proceed
• Normal operations
• No special precautions

Moderate Risk (0.25-0.5)

• Proceed with caution
• Monitor closely
• Prepare fallbacks

High Risk (0.5-0.75)

• Consider alternatives
• Activate safeguards
• Prepare for BBCR

Critical Risk (0.75-1.0)

• Immediate action required
• BBCR auto-trigger
• Return to safety

Monitoring

Continuous risk monitoring includes:

• Real-time ΔS tracking
• Logic coherence checks
• Confidence measurements
• Pattern detection
• Anomaly alerts

Integration

Risk assessment works with:

• `/boundary:detect` for position analysis
• `/wfgy:bbcr` for risk mitigation
• `/memory:checkpoint` for safe points
• `/reasoning:chain-validate` for logic checks