The Divine Shield
for Your Enterprise.

Institutional-grade quantitative risk management, real-time market data, and deterministic financial math—all inside your Claude chat interface.

aegis-v5.0 — zsh
vaibhavkkm@aegis ~ % /vkkm:ml-pd
[INFO] Running Scikit-Learn Logistic Regression Engine...
[INFO] Pulling financial ratios from Enterprise DB...

📊 PROBABILITY OF DEFAULT REPORT
------------------------------------------------
Predictor: ML-Logistic-Regression-v2
1-Year PD: 12.4%
Confidence Interval (95%): [10.2% - 14.6%]
Z-Score: 1.82 (Grey Zone)
------------------------------------------------
vaibhavkkm@aegis ~ % _

Institutional Capabilities

A direct bridge between LLM reasoning and professional quant systems.

Enterprise SQL

Direct integration with PostgreSQL, MySQL, and Snowflake. No file uploads needed—just live data.

Live Market Intelligence

Real-time volatility, closing prices, and T-Bill rates via yfinance. Grounded in today's reality.

Machine Learning PD

Calibrated logistic regression models for predicting default probability with rigorous math.

Visual Monte Carlo

In-chat simulations using Matplotlib. Visualize 1,000+ paths directly in your conversation.

FinBERT Sentiment

Local NLP models analyzing market news for quantitative bullish/bearish signals.

Board-Ready XLSX

One-click export of RAG-colored Excel risk registers and KRI dashboards.

Historical Backtesting

Proof-of-work: Kupiec POF tests to validate your Risk Models against the Basel Traffic Light system.

Local Zero-Egress

Institutional privacy by design. Monte Carlo models and SQL queries execute 100% locally.

The Command Center

18 institutional tools for full-spectrum risk oversight.

/vkkm:stress-test Simulate extreme market shocks across your portfolio.
/vkkm:var-calc Monte Carlo Value at Risk & Conditional VaR (ES).
/vkkm:greeks Black-Scholes calculation for Delta, Gamma, Vega, Theta.
/vkkm:backtest Kupiec POF test to validate your risk models.
/vkkm:plot-monte-carlo Generate visual path simulations in-chat.
/vkkm:sentiment-analysis FinBERT news sentiment scoring for market signals.
/vkkm:crypto-risk Volatile asset assessment via CoinGecko API.
/vkkm:liquidity-risk LCR, cash flow gap, and NSFR monitoring.
/vkkm:ml-pd ML-driven Probability of Default with 95% CI.
/vkkm:zscore Altman Z-Score bankruptcy predictor engine.
/vkkm:credit-risk Basel standard PD, EAD, LGD, and EL.
/vkkm:counterparty-profile Entity due diligence & risk profiling.
/vkkm:risk-register Generate ISO 31000 compliant risk registers.
/vkkm:rcsa Risk & Control Self-Assessment workflow.
/vkkm:kri-dashboard Track Key Risk Indicators (KRI) with RAG.
/vkkm:escalation-report Formalize internal risk escalation memos.
/vkkm:regulatory-check EU/Global compliance screen (GDPR, DORA, Basel).
/vkkm:icaap Structured 7-section ICAAP document generation.
/vkkm:reg-calendar 12-month rolling regulatory deadline tracking.
/vkkm:export-report Export risk data to Board-ready Excel files.
/vkkm:generate-pitchbook Generate a PDF Executive Summary to Desktop.
/vkkm:scan-document Scan contracts for 11 distinct risk patterns.

Autonomous Skills

Background intelligence that runs on every response.

Real-time Risk Scoring

Adds a 📊 AEGIS RISK SNAPSHOT (5x5 Matrix) to every risk-related response.

ALWAYS-ON

Pattern Intelligence

Scan-on-upload detection for liability, indemnification, and force majeure clauses.

AUTOMATIC

EU Regulatory Awareness

Flags GDPR, PSD2, DORA, Basel, NIS2, and EMIR exposure automatically.

COMPLIANCE

Plain-Language Bridge

Instantly translates technical quant findings for executive and board-level understanding.

REPORTING

Deploy Aegis in Minutes

Follow these simple steps to empower your Claude instance with institutional risk intelligence.

  1. Configure Claude Desktop

    Open your claude_desktop_config.json and add the Aegis MCP server configuration.

    "vkkm-aegis": {
      "command": "npx",
      "args": ["-y", "vkkm-aegis"]
    }
  2. Restart & Initialize

    Restart Claude and run your first command, for example /vkkm:stress-test, to initialize the local Python backend.

  3. Authenticate Databases

    Use the local environment variables to securely store your SQL connection strings for seamless integration.