Specboot is a CLI bootstrapper that installs a complete set of AI agent configurations, development standards, and reusable skills into any existing project. With a singleDocumentation Index
Fetch the complete documentation index at: https://mintlify.com/LIDR-academy/lidr-specboot/llms.txt
Use this file to discover all available pages before exploring further.
npx command it lays down a consistent, portable structure that every major AI coding copilot — Claude, Cursor, GitHub Copilot, and Gemini — can pick up immediately, without any manual wiring.
The Problem Specboot Solves
Modern development teams rarely use just one AI tool. A backend engineer might reach for Claude while a frontend engineer prefers Cursor, and another teammate uses GitHub Copilot inside VS Code. Without a shared configuration layer, each copilot operates in isolation: it has no knowledge of your architecture patterns, coding conventions, domain language, or API contracts. The result is inconsistent code quality, redundant context-setting in every session, and standards that live in people’s heads instead of in the repository. Specboot solves this by placing a single source of truth —docs/base-standards.md — at the center of your project and exposing it to every AI tool through the naming conventions each tool already understands.
Single Source of Truth
When you run Specboot, it creates four top-level configuration files as symbolic links, all pointing todocs/base-standards.md:
| File | Copilot |
|---|---|
CLAUDE.md | Claude / Cursor |
AGENTS.md | Generic (most copilots) |
codex.md | GitHub Copilot / Codex |
GEMINI.md | Google Gemini |
docs/base-standards.md once instantly propagates the change to every copilot. No drift, no duplication.
Key Components
Specboot installs three main areas into your project root:docs/ — Development standards and specifications
Contains base-standards.md (the single source of truth) alongside optional companion files: backend-standards.md, frontend-standards.md, documentation-standards.md, api-spec.yml, and data-model.md. These files define the rules every AI agent follows when generating code, tests, and documentation for your project.
ai-specs/agents/ — Agent role definitions
Markdown files that instruct an AI agent to adopt a specific role — for example backend-developer.md or frontend-developer.md. OpenSpec’s config.yml references these files to give each agent the right domain focus for the task at hand.
ai-specs/skills/ — Reusable skill prompts and workflows
Battle-tested prompt workflows packaged as SKILL.md files. Skills are auto-discovered by copilots when a request matches a skill’s description (governed by AGENTS.md §4). Highlights include enrich-us, using-git-worktrees, code-auditing, and writing-skills. All skills are mirrored into .claude/skills/ and .cursor/skills/ via relative symlinks so every supported tool can discover them without redundant copies.
The files Specboot installs in
docs/ are generic reference examples. Customizing them to match your real project — your stack, architecture patterns, domain language, API contracts, and data models — is mandatory. Without this step, your AI assistant will use generic context instead of your actual project context.Integration with OpenSpec
Specboot is designed to be used alongside OpenSpec, a spec-driven development framework. After bootstrapping, you point your project’sopenspec config.yml to the docs/ and ai-specs/ paths that Specboot created. From that point on, the full OpenSpec command flow — /enrich-us, /ff, /apply, /verify, /adversarial-review, /archive, /commit — has all the technical context it needs to generate accurate specs and implement them autonomously.
Who It’s For
Specboot is built for teams and individual developers who use AI coding assistants and want those assistants to produce consistent, high-quality code from day one. It is equally useful for:- Teams where engineers use different AI tools and need a shared set of standards
- Projects adopting spec-driven development with OpenSpec
- Developers who want reusable skill workflows without building them from scratch
- Organizations that need AI-generated code to follow real architecture patterns, not generic ones
Benefits
Multi-Copilot Support
CLAUDE.md, AGENTS.md, codex.md, and GEMINI.md are all symlinks to one source file. Switch copilots freely — the rules follow automatically.
Dev Standards
Core principles, TDD requirements, type safety, naming conventions, and language rules are documented in one place and applied by every agent on every task.
Agent Roles
Role-specific agent definitions in ai-specs/agents/ give each copilot the right domain focus — backend, frontend, analyst, and more — for the artifact it is creating.
Skills Library
Reusable SKILL.md workflows for common engineering tasks: story enrichment, git worktrees, code auditing, and more — auto-discovered by any supported copilot.