Documentation

Specification

The SpecScore format — how to write specs that lint clean.

• Overview — index of all specifications
• Feature — the atomic unit of product specification
• Plan — planning documents that bridge specs to execution
• Task — discrete units of work within a Plan
• Acceptance Criteria — verification conditions for requirements
• Repo Config — specscore.yaml configuration reference
• Source References — code-to-spec traceability
• Decision — Architecture Decision Records (ADRs)
• Decisions Index — canonical index of Decision documents
• Lessons — durable process-improvement rules with append-only occurrence evidence
• Writing and migrating Lessons — practical guide to compact rules, safe occurrence context, and lossless legacy migration
• Entities & Properties — the full entity model

Writing good specs

Craft guidance for turning requirements into clean specs.

• Asking Good Questions for Good Specs — how to question a fuzzy request into complete, unambiguous, verifiable acceptance criteria
• Explained by example: a login feature — one feature through the whole model, from Idea to Rehearse proof

Getting started

• Install the CLI — macOS, Linux, Windows
• CLI product page — what the linter does and why

AI plugins

• AI plugins overview — all SpecScore-compatible plugins
• SpecStudio Skills — opinionated SDD workflow for Claude Code
• CLI Skills — thin CLI wrapper for Claude Code

SpecScore for…

Role-specific guides for adopting SpecScore in your team.

• Developers
• Product Owners
• Business Analysts
• QA Engineers
• Project Managers
• Architects
• AI Agents

More

• Ecosystem — SpecScore Studio, Rehearse, Synchestra
• Rehearse — in plain language — the acceptance-evidence layer: what it does, how, and why
• Rehearse vs. Established Testing Frameworks — why not just use pytest or Cucumber?
• Blog — articles on spec-driven development

View as Markdown