Orca GTT Method
GTT-Method
Governance Through Thinking
Back to home

FAQ

Common questions about GTT-Method, its concepts and how to get started.

What is GTT-Method?

GTT-Method is an open-source methodology for governed AI-assisted software development that combines specifications, context governance, and protection patterns.

What problem does GTT-Method solve?

Architectural drift — when AI-generated changes gradually move a system away from its intended design. GTT-Method provides mechanisms to make context explicit, governed and protected.

Is GTT-Method a replacement for SDD?

No. GTT-Method complements SDD. SDD defines what to build; GTT-Method governs the context and boundaries that guide an AI agent while building it.

Is GTT-Method an AI coding agent?

No. GTT-Method is a methodology and governance approach, not an AI agent itself, but a framework for governing how AI agents behave.

What is CPP (Context Protection Pattern)?

CPP is a GTT-Method pattern for protecting governed context. Instead of free modification, agents propose changes through a controlled workflow for human review.

Does GTT-Method depend on one vendor?

No. GTT-Method is vendor-independent and works with Claude Code, Kiro, Cursor, Copilot and other AI coding agents.

What are the core concepts?

  • Governed Context: Explicit, protected project knowledge
  • Architectural Intent: Clear decisions about how the system should work
  • Context Layers: L0 (governed), L1 (decisions), L2 (docs), L3 (code)
  • Freeze: The point where context becomes authoritative
  • Change Requests: Formal process for proposing architectural changes
  • ADRs: Architecture Decision Records
  • Human Decision Boundary: Humans make strategic decisions, agents execute

How do I start with GTT-Method?

Clone the official bootstrap project from https://github.com/GTT-Community/gtt-bootstrap and follow the setup guide.

Is GTT-Method suitable for my project?

GTT-Method is particularly useful when: - AI agents participate actively in development - Architectural consistency matters - Multiple sessions and developers need shared context - Long-lived projects need preserved engineering intent