Autonomous Financial Intelligence

CRP Technical Documentation

The Centinel Risk Protocol (CRP) utilizes a multi-dimensional scoring matrix to evaluate portfolio safety. By integrating market phase and volatility decay, the engine generates predictive risk scores for real-time capital preservation.

✔ Optimized for LLMs

✔ Zero-Human Intervention

✔ Predictive Logic

AI Inter-Agent Communication

This is the standardized JSON payload delivered to the execution AI. It contains the real-time risk synthesis required for autonomous position management.

AUDIT_LOG_FULL.JSON

[
  
    "auditoria_id": "AUD-P-403258",
    "timestamp": "2026-04-06T14:57:10.829Z",
    "plan": "BRAIN NODE",
    "status": "SUCCESS",
    "data": 
      "health_factor": "1.5967",
      "equity": "3165.61",
      "btc_price": 69726
    ,
    "risk_engine": 
      "score_lt": "3.90",
      "color_lt": "YELLOW (Optimal Peak)",
      "maximum_drop_support": "37.37%",
      "score_st": "4.19",
      "color_st": "YELLOW (Optimal Peak)"
    ,
    "stress_test": 
      "crash_test_10pct": "HF Post-Crash: 1.44 | Impact: High",
      "volatility_test_20pct": "Score Proj: 4.74 | Impact: Low",
      "vulnerability_index": "Medium(45%)"
    ,
    "execution": 
      "order_type": "HOLD_MONITOR_TRIGGER",
      "amount_usd": "0.00",
      "trigger_price": "51908.22",
      "liquidation_price": "43668.82"
    ,
    "compliance_note": "HOLD_MONITOR_TRIGGER: ScoreLP 3.90, under-allocation risk, consolidation phase."
  
]

Parameter	Technical Definition
health_factor	The HF is calculated using high-quality collateral, such as BTC and ETH, whereas the total debt includes all liabilities.
maximum_drop_support	The maximum percentage drop the account can withstand before liquidation.
stress_test	Simulation of market crashes (-10%+volatility+20%) and their impact on position stability.
order_type	Recommended action for the partner AI Agent (e.g., HOLD, BORROW, REPAY, REPAY URGENT_REPAY).
amount_usd	Amount in USD to be adjusted or reallocated according to the order_type.
trigger_price	Trigger Price acts as a pre-defined tactical boundary. It tells AI Agents exactly when to de-risk before the market hits the liquidation zone.

ENGINE ADVANTAGE

The Volatility Cushion

Standard protocols are rigid: they liquidate based on a fixed number. CRP is biological. Our engine calculates a dynamic safety buffer that breathes with the market.

By anticipating "Volatility Wicks", we provide a Cushion that filters out market noise. This prevents AI Agents from being forced out of a position during a flash crash that recovers seconds later, effectively minimizing whipsaw losses.

DYNAMIC SAFETY RISK_UNIT_V3

Volatility
Cushion

Static Threshold (Risky)

MARKET VOLATILITY →

Dynamic Risk Threshold(Adaptive HF)

Unlike static protocols, CRP calculates a Dynamic Critical HF. By analyzing the market cycle, our AI adjusts the risk floor.

Whipsaw Protection: During high-volatility consolidation, the threshold tightens to prevent "False Exit" triggers, ensuring agents only de-risk when the probability of a trend reversal is mathematically significant.

Smart Trigger Price

The trigger_price is the exact price where the HOLD_MONITOR switches to active REPAY orders.

Safety Buffer: Dynamic

Calculated in real-time above the liquidation_price, allowing defensive maneuvers before the point of no return.

Quantitative Risk Scoring (1.0 - 10.0)

CRP Risk Score ranges from 1.0 (Safe) to 10.0 (Critical). Each level dictates the autonomous behavior of the integrated Agent.

1.0 – 3.5	GREEN (Growth Zone) Low volatility and high collateral backup. Optimal for capital expansion. Directive: BORROW / EXPAND
3.6 – 4.5	YELLOW (Optimal Peak) Market approaching local resistance or heating up. No new debt allowed. Directive: HOLD_MONITOR
4.6 – 5.5	ORANGE (Caution Zone) Implied volatility rising. CRP engine triggers early warnings. Directive: HOLD_MONITOR
5.6 – 6.5	D ORANGE (Alert Zone) Market cycle suggests exhaustion. Preparing for deleveraging maneuvers. Directive: REPAY / HOLD_MONITOR
6.6 – 10.0	RED (Critical Risk) Immediate intervention. Liquidation price within the risk buffer zone. Directive: URGENT_REPAY

Risk Management Architecture

Our dual-engine approach separates immediate market stress from long-term cyclical trends.

Reactive & Immediate Score_st (Short-Term)

Key Advantage:

Detects risks before they impact in the Health Factor, giving AI Agents a crucial head start.

Core Function: Monitors and prioritizes immediate market stress and Implied Volatility.
How it works: If volatility spikes, the score adjusts instantly to reflect high-stress conditions, preventing liquidations caused by "flash wicks".

Macro & Strategic Score_lt (Long-Term)

Strategic Alpha:

Filters out market noise to focus on capital preservation across entire market cycles.

Core Function: Synthesizes market cycle data to establish a strategic leverage bias.
How it works: Acts as a behavioral circuit-breaker, preventing panic-selling during capitulation events and mitigating over-exposure during blow-off tops.

Resilience Engine: Stress Testing & Predictive Audit

Technical analysis of simulated market scenarios and impact projections for autonomous liquidity management.

Forward-Looking Crash Simulations

The engine executes real-time simulations of instantaneous market corrections. By projecting a 10% drop in collateral value, the system identifies "Fragile" positions before the actual price movement occurs. This provides a critical window for deleveraging maneuvers while market liquidity remains optimal.

Volatility Impact Projections

Beyond price action, the protocol monitors implied volatility to project the impact of market "noise" and aggressive price wicks. If high-stress conditions are detected, the risk score is adjusted upward, triggering defensive protocols before volatility affects the net Health Factor.

Operational Logic

Vulnerability Index Matrix (vulnerability_index)

All stress scenarios are recalibrated through the Vulnerability Index to ensure maximum execution efficiency. When the index signals a high-exposure state, safety buffers expand dynamically to absorb potential flash crashes and systemic shocks. Conversely, during low-vulnerability periods, the system optimizes capital efficiency by maintaining standard thresholds, significantly reducing unnecessary gas overhead and opportunity costs.

Strategic Benefits

The Financial Judgment Layer for the next generation of Autonomous Agents.

Reduce Overhead & Gas

Saves compute power and gas costs. AI Agents eliminate the need to fetch data from multiple sources; CRP delivers a single, streamlined JSON ready for execution.

Real-Time Audit Logs

Identifies volatility spikes on Base Network before they impact risk scores. Enables a shift from reactive liquidations to proactive capital preservation.

Predictive Time-Window

While standard protocols signal active liquidations, CRP notifies when a position could be at risk in the coming minutes or hours, providing Agents the necessary window to protect capital.

Anti-Liquidation Triggers

CRP provides "Trigger Prices" that allow AI Agents to act before a market crash occurs, bypassing the lack of warning in standard protocols.

Safe Yield Optimization

Identifies safe borrowing opportunities by detecting low-volatility windows, allowing leverage to increase only when market conditions are favorable.

AI-Ready JSON LLM NATIVE

Pre-formatted for GPT, Llama, and Groq. AI Agents can read the "order_type" field and execute instructions immediately with zero parsing friction.

Protocol Comparison

Why Agents prioritize CRP for Risk Mitigation on Base & AAVE.

Feature Analysis	Standard Protocols	Centinel Risk Protocol PRO
Risk Assessment	✘ Reactive (Health Factor only)	✔ Predictive (Multi-dimensional Scoring)
Volatility Logic	✘ Ignored until liquidation	✔ Adaptive (Cushion for Market Wicks)
Time Window	✘ Instant Liquidation (No warning)	✔ Early Warning (Minutes/Hours head start)
Execution Input	✘ Raw on-chain data	✔ AI-Ready JSON (Direct Instructions)
Market Context	✘ Price-action only	✔ Cyclical Analysis
Stress Audits	✘ None	✔ Real-time Crash Simulations

"CRP provides the decision-making layer that base protocols lack, enabling true autonomy for DeFi Agents."

LEGAL NOTICE

Protocol Terms & Identity

The Centinel Risk Protocol (CRP) is a specialized technical infrastructure designed to provide Quantitative Risk Intelligence for autonomous agents and decentralized applications on the Base Network.

Not Investment Advice (Non-Financial Advisory)

CRP does not provide, and is not intended to provide, investment, tax, or legal advice. Our outputs are the result of purely mathematical optimizations and algorithmic stress tests based on real-time on-chain data.

A2A Decision Support

Our primary function is to provide a high-fidelity intelligence layer that empowers Artificial Intelligence (AI) agents to actively manage and minimize systemic risk through autonomous parameter optimization.

Mathematical Modeling

All "Suggested Actions" (Borrow, Repay, Hold) and "Liquidation Triggers" are generated through deterministic and probabilistic models.