sparc_methodology

Orchestrates multi-agent software development using the SPARC specification-to-completion methodology.

Updated Jan 14, 2026
One-click install
npx skills add https://github.com/jvsandhu/agentic-skills --skill sparc-methodology-jvsandhu
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sparc_methodology
Source: https://github.com/jvsandhu/agentic-skills/tree/main/skills/sparc_methodology
Command: npx skills add https://github.com/jvsandhu/agentic-skills --skill sparc-methodology-jvsandhu

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? It removes the chaos from complex software projects by enforcing a structured Specification, Pseudocode, Architecture, Refinement, and Completion workflow with coordinated AI agents instead of ad-hoc coding. ## Core Features & Use Cases - 17 Specialized Modes: Orchestrator, architect, coder, tdd, reviewer, researcher, debugger, and more, each handling a distinct development phase. - Multi-Agent Orchestration: Hierarchical, mesh, and pipeline swarm topologies with parallel execution for 2.8-4.4x speed improvements. - TDD Enforcement: Red-green-refactor cycles with 90%+ coverage targets and memory-based cross-agent coordination. - Use Case: Building a full-stack authentication feature by running researcher, architect, coder, tdd, and reviewer modes in a coordinated pipeline via claude-flow MCP tools or NPX CLI. ## Quick Start Ask the agent to run the SPARC pipeline for your feature, for example: use the sparc methodology to design and implement a JWT authentication API with tests.

Frequently Asked Questions about sparc_methodology

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

FAQPage Schema
How do I run a SPARC mode with claude-flow?

Use the MCP tool mcp__claude-flow__sparc_mode with a mode name and task description, or run npx claude-flow sparc run <mode> "task" from the terminal. You can also execute a full pipeline with npx claude-flow sparc pipeline "task description".

What are the phases of the SPARC methodology?

SPARC has five phases: Specification (requirements and constraints), Architecture (system and component design), Refinement (TDD implementation), Review (quality, security, performance checks), and Completion (integration, deployment, documentation).

Which swarm topology should I use for multi-agent orchestration?

Use hierarchical topology for complex projects with clear delegation, mesh for peer-to-peer collaborative tasks, and sequential pipelines for ordered spec-to-review workflows. Adaptive strategy auto-adjusts for dynamic workloads.

Does SPARC support test-driven development workflows?

Yes, SPARC includes a dedicated tdd mode enforcing the red-green-refactor cycle with a 90% coverage target. It supports Jest, Mocha, Vitest for unit tests and Playwright or Cypress for end-to-end testing.

What are the limitations of the SPARC multi-agent approach?

SPARC depends on claude-flow being available via MCP tools or NPX CLI, so it cannot run standalone. Coordination overhead may not pay off for small single-file tasks that do not need multi-phase orchestration.