speckit

An AI-assisted coding workflow with governance gates uses a pipeline router to sequence work through specify, research, plan, implement, test, e2e, and retro phases. It verifies compliance constraints before each transition and auto-continues on successful execution to streamline development.

Living docs and issue-backed specs are maintained by routing research and retro states through a structured pipeline. This integration ensures documentation stays synchronized with project phases, preserving traceability from specification through implementation and testing.

A pipeline router for AI-coding workflows supports both greenfield and brownfield projects. It adapts the workflow by routing tasks based on the current project phase, applying governance checks and extension hooks regardless of the existing codebase maturity.

Traceability across AI coding pipelines is ensured by routing work through issue-backed steps with governance checks before each transition. This enforces compliance and maintains a verifiable record from initial specification through end-to-end testing and retro.

A development pipeline can be extended with custom sub-skills and subagents by leveraging its extension hooks. These hooks allow you to inject specialized routing logic or governance constraints, adapting the specify through retro workflow to specific project needs.

An automated governance pipeline for AI coding may be unsuitable for highly exploratory or ad-hoc tasks that do not benefit from structured phase routing. If your work does not require issue-backed traceability or compliance checks, the overhead of specify through retro steps may be unnecessary.