sakura-dev

Leads software development from requirements through decomposition, execution, verification, and delivery.

3|1|Updated Sep 9, 2026
One-click install
npx skills add https://github.com/v20227/using-sakura-dev --skill sakura-dev-v20227
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sakura-dev
Source: https://github.com/v20227/using-sakura-dev/tree/main/skills/sakura-dev
Command: npx skills add https://github.com/v20227/using-sakura-dev --skill sakura-dev-v20227

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Multi-step software work often loses continuity: requirements drift, task dependencies get forgotten, fixes overwrite earlier behavior, and completion is declared without evidence. This Skill maintains one recoverable development workstream so a project moves from a natural-language goal to a verified, deliverable result without losing prior decisions. ## Core Features & Use Cases - Primary workflow orchestration: Recovers existing project state, maintains scope and acceptance conditions, decomposes work into ordered items, executes, verifies, and closes each milestone with recorded evidence. - Continuity and repair history: Preserves accepted behavior, failed attempts, and the exact next action across contexts, reopening items when regressions disprove earlier acceptance. - Specialist Skill coordination: Directs supporting Skills for specification, planning, coding, verification, fixing, review, release, research, delegation, simplification, and security auditing within their own boundaries. - Use Case: You ask to add a multi-module feature to an existing app. The Skill recovers the current project records, updates the scope and acceptance conditions, splits the work into dependency-ordered items, implements and verifies each one against the real user path, and delivers a runnable result with its launch entrypoint. ## Quick Start Tell the assistant to continue or start your software project using the sakura-dev workflow and describe the outcome you want in plain language.

Frequently Asked Questions about sakura-dev

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I manage a multi-step software project with an AI coding assistant?▼

Describe the outcome in plain language and let the workflow recover existing project records, decompose the work into dependency-ordered items, and execute them in sequence. Each item carries acceptance conditions and verification evidence, so progress survives context switches and new messages.

How do I keep requirements and progress from being lost between conversations?▼

Keep scope, work items, acceptance, evidence, and the exact next action in the project's .sakura/ records or an established equivalent. On resume, the workflow reads those records, reconciles them with the actual source, and continues from the recorded next action.

When should I use a full development workflow instead of a quick code edit?▼

Use the full workflow for work with multiple meaningful changes, connected boundaries, or repeated corrections. A genuinely independent small change stays a compact work item with its check and result, and read-only questions stay read-only.

Does this workflow replace the AI's native coding and terminal tools?▼

No. It organizes native capabilities such as file editing, search, command execution, browser use, and subagents into a verifiable development mainline. The assistant still selects and uses the actual tools available in the environment.

What counts as proof that a feature is complete?▼

Completion requires each acceptance condition verified against the tested source or artifact, including the actual user path and affected preserved behavior. Passing an existing test suite, reading a Skill, or receiving a subagent result alone does not satisfy acceptance.

What are the limitations of this development workflow?▼

It depends on the host environment's available tools and permissions, and it does not invent missing capabilities such as subagents or persistent goals. Materially ambiguous product direction still requires a user decision before implementation proceeds.