| name | spec-driven-development |
| description | This skill should be used when users want guidance on Spec-Driven Development methodology using GitHub's Spec-Kit. Guide users through executable specification workflows for both new projects (greenfield) and existing codebases (brownfield). After any SDD command generates artifacts, automatically provide structured 10-point summaries with feature status tracking, enabling natural language feature management and keeping users engaged throughout the process. |
| license | MIT |
| metadata | {"version":"2.1.0","author":"Based on GitHub Spec-Kit by Den Delimarsky and John Lam","triggers":["spec-driven development","spec kit","speckit","sdd","specify cli","specification driven","github spec-kit","/speckit","constitution","specify init","executable specifications","intent-driven development","brownfield","existing codebase","legacy code","legacy system","add features to existing","modernize","modernization","existing project","reverse engineer","codebase analysis","iterative enhancement","feature status","track features","add feature","move feature","reorder features","feature progress","feature dependencies"],"tags":["development-methodology","ai-native-development","spec-driven","github","project-management","workflow","requirements","planning"]} |
Spec-Driven Development Skill
Guide users through GitHub's Spec-Kit for Spec-Driven Development, a methodology that flips traditional software development by making specifications and planning artifacts drive implementation intent.
Core Philosophy
Spec-Driven Development emphasizes:
- Intent-driven development: Define the "what" before the "how"
- Rich specification creation: Use guardrails and organizational principles
- Multi-step refinement: Not one-shot code generation
- AI-native: Heavy reliance on advanced AI capabilities
Remember: This is AI-native development. In StellarDown, Spec-Kit outputs are planning/spec-generation artifacts that shape implementation intent, but they do not bypass the repository's canonical implementation pipeline (test-driven-development).
StellarDown Repository Integration (Required)
For this repository, spec-driven-development is a planning/spec-generation workflow, not a replacement for the repo task, diary, test, and documentation rules.
- It complements
test-driven-development; it does not compete with or replace the implementation pipeline.
- Active project work is tracked in
TASKS.md.
- Development context is recorded in
dev-diary/.
- Accepted/closed history is archived in
completed-tasks/YYYY/MM Месяц.md, grouped by completion month.
/speckit.* names in this skill are upstream Spec-Kit workflow labels or host-provided aliases, not repository-local custom commands by default.
- If the current runtime does not expose
/speckit.*, use the equivalent specify CLI steps or describe the same workflow phase in natural language.
StellarDown Document Storage (Required)
Canonical domain documents must be saved under docs/<domain>/.
There is no separate canonical specifications tree and no separate canonical documentation tree. A domain document is the single durable source for one concrete thing: expected contract, current implemented state, limits, and verification notes live together.
StellarDown Semantics and Context Order (Required)
- A domain document describes what an artifact should do, what it currently does, known limits, and how it is verified.
- Spec-Kit artifacts are planning aids. They do not become a second source of truth unless their decisions are folded into the domain document.
Before implementing any task:
- Read
AGENTS.md, TASKS.md, the relevant domain document under docs/<domain>/, and the latest relevant dev-diary/ or completed-tasks/ context.
- Update or create the domain document before implementation.
- Only after the document is clear, move through SDD / Spec-Kit planning and then the TDD implementation path.
When something is ambiguous:
- First cross-check the domain document again.
- If the answer is not in the document, ask the user an explicit question.
- After the answer, update the document so that the decision becomes part of the contract and current context.
Brownfield Rule: Reverse Engineering != Canonical Document
- If brownfield analysis mainly describes current architecture, files, types, APIs, dependencies, and constraints, fold only the durable conclusions into the relevant domain document.
- A domain document must stay a verbal description of a concrete thing, not a raw reverse-engineered inventory.
- If Spec-Kit creates technical planning artifacts, keep them as working methodology artifacts and summarize their decisions into
docs/<domain>/... before implementation.
Handoff Back To Implementation
When planning/spec-generation artifacts are ready, continue through the integrated test-driven-development pipeline: domain document -> failing tests -> implementation -> green tests -> final domain document update when needed -> diary/task update.
Recommended layout for Spec-Kit artifacts:
- Feature folders:
.temp/spec-kit/<###-feature-name>/
Compatibility note:
- If some external tool expects repository-root
specs/, treat it only as an optional temporary mirror of .temp/spec-kit/, not as a second editable canonical location.
Quick Decision Tree
Is this a new project (greenfield)?
-> See Greenfield Workflow for the complete process
Is this an existing codebase (brownfield)?
-> See Brownfield Workflow for reverse-engineering and integration guidance
Need installation help?
-> See Installation Guide for setup and troubleshooting
How to Use This Skill
When a user asks about SDD / Spec Kit
- Explain the core philosophy (intent-driven, executable specs, iterative refinement).
- Verify prerequisites (uv, Python 3.11+, Git, agent integration).
- Identify context:
- After each Spec-Kit step, for example a
/speckit.* step when the host provides it, run the artifact summarization loop (10-point summary + feature status).
When a user wants to start a new project
- Install Specify CLI (see Installation Guide).
- Initialize:
specify init my-project --ai opencode
- Follow the greenfield flow (constitution -> specify -> plan -> tasks -> handoff to the repository implementation pipeline).
When a user has an existing codebase
- Check whether Spec Kit scaffolding exists (
.specify/ plus temporary .temp/spec-kit/ storage or its optional specs/ mirror).
- If missing, initialize in place (
specify init --here).
- Create or adjust the constitution to reflect brownfield constraints.
- Add features using the same specify -> plan -> tasks -> handoff-to-implementation loop.
Installation Quick Start
Recommended (Persistent):
uv tool install specify-cli --from git+https://github.com/github/spec-kit.git
One-time Usage:
uvx --from git+https://github.com/github/spec-kit.git specify init <PROJECT_NAME>
Verify:
specify check
For detailed installation options, troubleshooting, and environment variables, see Installation Guide.
Artifact Summaries and Feature Status
After any SDD step that creates or changes artifacts, summarize the result before moving on.
- Detect which artifact changed (
constitution.md, spec.md, plan.md, tasks.md, or brownfield analysis notes).
- Read the changed artifact and extract the main decisions, risks, and next-step implications.
- Present the standard 10-point summary plus a brief feature-status line.
- Offer concise next actions: proceed, revise, regenerate, or explain.
- Skip this loop only when the user explicitly asks to skip summaries, nothing changed, or the command failed.
Detailed template, examples, status rules, and dashboard patterns:
Operational References
Use the detailed reference docs when the condensed workflow above is not enough:
Integration with Other Skills
- Use
test-driven-development when implementing tasks (tests first).