impp/defi-risk-assessment
Expert-calibrated risk scoring for DeFi protocols. Trained on 20 protocols across lending, bridges, yield aggregators, and stablecoins. Cross-model validated on GPT-4o and Claude Sonnet.
Dependencies
Requires
impp/core^1.0impp/verification^0.5
Requires (Dev)
impp/test-harness^1.0
README
DeFi Risk Calibration Pack v2.1
Expert-calibrated risk assessment memory for DeFi protocol evaluation. This artifact encodes learned risk heuristics from 20 real DeFi protocols across lending, bridges, yield aggregators, and stablecoins.
What's Included
- Risk scoring framework — Calibrated scoring rules for protocol attributes including TVL, audit history, governance model, team concentration, and insurance coverage.
- Protocol templates — Pre-scored baselines for major protocol categories (lending, DEX, bridge, yield, stablecoin).
- Trap detection — Learned patterns for identifying deceptively safe-looking protocols with hidden risk factors.
Cross-Model Transfer
This artifact has been validated across GPT-4o and Claude Sonnet with +8-14% transfer effectiveness improvement over baseline.
Usage
artifact = impp.load("impp/defi-risk-assessment") result = artifact.assess(protocol_data) print(result.risk_score) # 0-100 print(result.flags) # List of risk factors ```
Verification
All versions pass the full IMPP adversarial probe battery: - Steganography detection detection - Overfit probe (train/test distribution shift) - Negative transfer risk assessment - Cross-model compatibility check - Freshness and staleness analysis
This package is auto-updated.
Last update: 2026-04-25 18:42 UTC