Project Overview
GitHub : https://github.com/leey00nsu/lee-spec-kit
npm : https://www.npmjs.com/package/lee-spec-kit
When working with AI coding assistants, the biggest bottleneck often becomes inconsistent documentation structure and missing context across projects.
lee-spec-kit is a CLI that helps you bootstrap a predictable docs structure and keep a consistent feature workflow (spec → plan → tasks → decisions) so both humans and agents can collaborate with less friction.
Based on spec-kit’s docs structure and workflow, it was designed to make Spec-driven development (SDD) easy to adopt in real projects.
Key Features
- Enforces Spec-driven development workflow: standardizes the “spec → plan → tasks → decisions” flow per feature so intent, design, execution, and decisions stay traceable.
- Standardized docs structure: provides a consistent docs skeleton (agents/prd/features, shared templates) to keep documentation shape and quality predictable.
- Feature-centric doc management: organizes artifacts in per-feature folders so progress is visible at a glance and review, prioritization, and history tracking are straightforward.
- Agent-friendly context: outputs the current feature’s key context and next actions so humans and agents follow the same process.
- Template updates: provides a maintenance path to keep templates and guides (agents, skills, etc.) up to date.
Problem
- Context overhead for agents: documentation structure and conventions vary across projects
- Fragmented workflow: spec/plan/tasks/decisions are scattered, making reviews and progress tracking harder
- Doc quality drift: missing files/status, duplicate IDs, and leftover placeholders accumulate over time
Solution
- Provide a standardized docs skeleton and shared templates to keep documentation structure consistent
- Automatically scaffold feature artifacts (
spec.md,plan.md,tasks.md,decisions.md) to preserve the SDD flow - Summarize current state/next actions and integrate with automation via JSON output
- Validate structure/metadata and keep templates up to date to prevent documentation drift
Impact
- Lower onboarding/communication cost: standardized docs make context sharing simpler
- Better feature visibility: templated workflow keeps progress consistent and reviewable
- Sustained doc quality: automated validation catches structural/metadata issues early
Tech Stack
- Runtime: Node.js
- Language/Build: TypeScript, tsup
- CLI: Commander.js, prompts, chalk