Spec Kit

Spec Kit gives developers structured processes for working with AI coding agents. It turns requirements, assessments and verification results into project documents you can review alongside the code.

• Specification-led development: Carry requirements through a specification, technical plan and tasks, then guide implementation and check its fit against the specification.
• Bug diagnosis and repair: An optional extension separates cause assessment, a scoped repair and checks of the original symptom, with a recorded verdict of verified, partial or failed.
• Idea assessment: Another optional extension gathers evidence and records a decision to proceed, seek clarification or stop, even in a project without source code.
• Custom processes: Adapt reusable templates and behavior, add capabilities through extensions, or bring your own process; workflows automate steps and bundles group resources for particular roles.

Specification-led development is part of the core toolkit. Bug fixing and idea assessment are independent processes, so you don't need to complete a feature workflow before using them. Assessment records are Markdown documents you can refine as questions get resolved.

Spec Kit runs locally on Linux, macOS and Windows and is open source under the MIT license. It requires Python 3.11+, uv and a separate supported AI coding agent, including GitHub Copilot.