MarcelRoozekrans 

Async-first interfaces and base classes for ADO.NET (System.Data). Drop-in replacement with modern async/await, IAsyncEnumerable, and ValueTask support.

Async DataSet and DataTable for ADO.NET (System.Data). Includes AsyncDataTable, AsyncDataSet, and AsyncDataAdapter.

Roslyn-based MCP server that gives AI agents deep semantic understanding of .NET codebases — navigation, call graphs, diagnostics, refactoring, code-quality auditing, test intelligence, and IL/external-assembly... More information

Adapters wrapping existing ADO.NET providers (DbConnection, DbCommand, DbDataReader) into async interfaces. Includes .AsAsync() extension and DI support.

Newtonsoft.Json converters for AsyncDataTable and AsyncDataSet, wire-compatible with Json.Net.DataSetConverters.

System.Text.Json converters for AsyncDataTable and AsyncDataSet, wire-compatible with Json.Net.DataSetConverters.

Roslyn source generator that produces typed async DataSet/DataTable/DataRow classes from .xsd schema files.

Security monitoring middleware for IChatClient — prompt injection, hallucination, and operational anomaly detection with an intervention engine.

Claude Code hook adapter for AI.Sentinel — scan UserPromptSubmit, PreToolUse, and PostToolUse hook payloads through the detector pipeline.

Claude Code hook command for AI.Sentinel — wire into settings.json hooks to scan tool calls.

MCP server for .NET memory profiling — wraps JetBrains dotnet-dotmemory with heuristic-based analysis rules

MCP proxy CLI for AI.Sentinel — point your MCP host at sentinel-mcp to scan all tool calls.

GitHub Copilot hook command for AI.Sentinel — wire into hooks.json to scan tool calls.

Embedded ASP.NET Core dashboard for AI.Sentinel — real-time threat monitoring, audit trail, and agent graph.

GitHub Copilot hook adapter for AI.Sentinel — scan userPromptSubmitted, preToolUse, postToolUse hook payloads through the detector pipeline.

Forward AI.Sentinel audit entries as OpenTelemetry log records — vendor-neutral.

Offline replay CLI for AI.Sentinel — run the detector pipeline against saved conversations.

Forward AI.Sentinel audit entries to Azure Sentinel via the Logs Ingestion API.

MCP (Model Context Protocol) proxy for AI.Sentinel — wraps another MCP server and scans tool calls in both directions.

Persistent SQLite-backed audit store for AI.Sentinel — survives restarts, maintains the hash chain, supports time-based retention.