A working demonstration of a Level 3 agentic SDLC on a real .NET codebase — spec-driven development, AI code and test generation, self-healing build loops, and human-in-the-loop review gates.
- Spec-driven development — features are specified with acceptance criteria and edge cases before any code is written
- Agentic code + test generation — AI agents (Claude Code, GitHub Copilot) implement features from specs on dedicated feature branches
- Self-healing loop — agents build, test, and auto-correct failures before handing off for review
- Human control points — every feature goes through a human merge gate
- Simulated MCP integration — a notification gateway shows how a real MCP connection (ADO, Confluence) would plug into the workflow
| File | What it is |
|---|---|
CLAUDE.md |
Agent operating manual — standards, routing rules, self-healing loop |
AGENTS.md |
Sub-agent roles and how they hand off work |
docs/SDD_TASKMANAGEMENT.md |
Feature specs, written for agents to implement directly |
src/ |
The .NET 8 Web API (Core → Infrastructure → Api layers) |
tests/ |
xUnit tests, generated alongside each feature |
.NET 8 · ASP.NET Core Web API · EF Core (SQLite) · xUnit · Moq
dotnet build
dotnet test
dotnet run --project src/TaskManagement.Api