TchMind MetadataAgent | Sentinel Core

💡 Overview

TchMind MetadataAgent is a deterministic, fault-tolerant execution engine designed for high-integrity metadata inspection and code generation within the DataHub ecosystem. Built for production-grade environments, it ensures that every transformation artifact (DAG) is generated with structural integrity, lineage awareness, and cryptographic verification.

🏗️ Architecture: Physical-Deterministic Engine

Our system follows a "closed-loop" operational philosophy:

• Deterministic Inspection: Strict schema validation ensuring zero-drift metadata interpretation.
• Integrity Verification: Every generated artifact is signed with a SHA-256 checksum to ensure consistency.
• Fault-Tolerant Loop: Automatic retry mechanisms prevent system failure during transient network or schema instability.

🚀 Key Capabilities

• Lineage-Aware Generation: Automatically pulls connection data from DataHub to build production-ready Airflow DAGs.
• Safety Assertion: Fails fast if schema corruption is detected to prevent downstream deployment errors.
• Retry Logic: Implements an exponential backoff strategy (configurable) to handle distributed system volatility.

🛠️ Tech Stack

• Engine: Python 3.10+
• Security: SHA-256 Checksum Validation, Assert-based schema inspection.
• Platform: Integrated with DataHub Context Kit for real-time metadata access.

⚙️ Quick Start

1. Configure your DataHub credentials in keys/vertex-key.json.
2. Initialize the Sentinel Core Agent:

from src.agents.MetadataAgent import DataHubMetaDataAgent

# Initialize with your DataHub context client
agent = DataHubMetaDataAgent(context_kit=YourDataHubClient())

# Execute safe, lineage-aware task generation
result = agent.execute_task("urn:your:dataset")

print(f"Execution Status: {result['status']}")
print(f"Artifact Checksum: {result.get('checksum')}")

Built With

• ai-agents
• apache-airflow
• datahub
• git
• google-cloud
• llm
• mcp
• metadata-api
• python
• typescript
• vertex-ai