ai-development-loop

Ticket-driven development is automated through a 9-step loop covering foundation checks, ticket selection, implementation, verification, and reflection to enforce consistent AI project quality. It integrates with the ticketr CLI to manage atomic ticket updates.

The AI development loop is a 9-step cycle: foundation check, ticket selection, start, high quality, strategy, implementation, verification, ticket audit, completion, and commit. It enforces quality gates and tracks reflection opportunities for continuous project improvement.

Quality gates are enforced during the development cycle by requiring verification and ticket audits before completion. The workflow integrates environment-aware execution and atomic ticket updates to maintain standards across foundation, implementation, and reflection phases.

Yes, the ticketr CLI is required as the workflow integrates directly with it for ticket selection and auto-generated templates. The loop uses tkr ready to prepare the next ticket and manages atomic ticket updates on start and completion.

AI agent development cycles are standardized through environment-aware execution supporting multiple managers. The workflow adapts to your configured environment using dependencies like docker, direnv, devbox, mise, nix, poetry, and uv for consistent project execution.

Reflections and learnings are tracked through a formal reflection and opportunity-tracking process built into the development loop. After verification and completion, the workflow documents learnings to improve future ticket-driven cycles.