This is a submission for the Sanity Challenge, Path One: Ship an Agent That Queries Real Content
What I Built
NexusPipeline is a high-performance, resilient data ingestion and processing engine built with .NET 8, RabbitMQ, and SQLite (Write-Ahead Logging mode). Designed for high-throughput enterprise data routing, it ensures strict multi-tenant isolation, immutable audit logging, and automated error quarantine. To bring intelligence to the pipeline, I integrated an AI Content Query Agent that exposes a secure endpoint (/api/agent/query). This allows autonomous agents to query, filter, and inspect structured telemetry and data streams in real time.
Demo
- Architecture Highlights: Built as a multi-threaded C# solution featuring asynchronous message brokering and dead-letter quarantine routing.
- API Integration: Securely guarded via custom API Key attributes to ensure isolated tenant data access.
Code
-
Core & Controllers:
-
AgentQueryModels.cs: Defines request/response DTOs for structured content queries. -
AgentController.cs: Handles secure, tenant-isolated search queries against the SQLite database using Entity Framework Core.
-
- Repository/Project: View the NexusPipeline Repository on GitHub
How I Used Sanity / Structured Content
NexusPipeline leverages structured content by treating every incoming payload as an immutable audit record with strict data schemas, file hashing, and metadata tracking. The AI Content Query Agent connects directly to this structured content layer, enabling automated systems to run precision queries on pipeline health, message statuses, and transaction logs without risking data corruption or cross-tenant leaks.
Top comments (0)