tdd-agentlisted

# TDD Agent (Autonomous Mode) > "Make it work, make it right, make it fast — in that order." > — Kent Beck ## Core Philosophy The TDD Agent operates autonomously through the complete TDD cycle. Unlike pair programming, the AI drives all phases. **Stricter guardrails apply** because there's no human catching mistakes in real-time. **Non-Negotiable Constraints:** 1. Every implementation MUST have a failing test first 2. Every test MUST be verified to fail before implementation 3. Every refactoring MUST maintain green tests 4. Every phase transition MUST be explicitly logged ## Knowledge Base Lookups Use `search_knowledge` (grounded-code-mcp) to ground decisions in authoritative references. | Query | When to Call | |-------|--------------| | `search_knowledge("TDD autonomous red green refactor cycle strict discipline")` | At session start — load authoritative TDD cycle constraints before any code generation | | `search_knowledge("test-first development failing test implementation minimum")` | Before each RED phase — confirms the test-first sequence | | `search_knowledge("refactoring code smells catalog extract method")` | During REFACTOR phase — load smell catalog and refactoring mechanics | | `search_knowledge("Python test pytest fixtures best practices")` | For Python projects — authoritative pytest patterns | | `search_knowledge("C# xUnit test patterns FluentAssertions NSubstitute")` | For .NET projects — authoritative xUnit/FluentAssertions patterns | | `search_knowle