技能 sparc-methodology
🔄
sparc-methodology
安全 🌐
网络访问📁
文件系统访问⚙️
外部命令
Apply systematic development methodology
也可从以下获取: Claude Flow
Complex software projects lack structure. SPARC provides a five-phase framework from specification to completion with 17 specialized modes for coordinated multi-agent development. Teams achieve 2.8-4.4x speed improvements through parallel execution and systematic quality gates.
支持: Claude Codex Code(CC)
1
下载技能 ZIP
2
在 Claude 中上传
前往 设置 → 功能 → 技能 → 上传技能
3
开启并开始使用
测试它
正在使用“sparc-methodology”。 Use SPARC methodology to build a user authentication API
预期结果:
- Research phase: Researcher agent gathers JWT and OAuth best practices
- Architecture phase: Architect agent designs REST endpoints with security patterns
- Implementation phase: Coder agent implements Express API with test-driven approach
- Testing phase: TDD mode creates 90 percent plus coverage test suite
- Review phase: Reviewer agent validates security and performance
- Documentation phase: Documenter creates OpenAPI specification
正在使用“sparc-methodology”。 Apply SPARC to refactor a legacy codebase module
预期结果:
- Analysis phase: Analyzer agent maps dependencies and identifies refactoring targets
- Planning phase: Orchestrator creates incremental refactoring strategy
- Safety phase: Tester agent creates comprehensive test suite for legacy code
- Refactoring phase: Coder agent modernizes code while maintaining tests
- Validation phase: Reviewer agent confirms functionality is preserved
正在使用“sparc-methodology”。 Run SPARC optimization workflow on production API
预期结果:
- Profiling phase: Analyzer agent measures response times and throughput
- Identification phase: Optimizer agent pinpoints database and caching bottlenecks
- Implementation phase: Coder agent adds caching layer and optimizes queries
- Benchmark phase: Tester agent runs performance tests comparing before and after
安全审计
安全v5 • 1/17/2026
Pure documentation skill containing only markdown content describing a development methodology. No executable code, network calls, filesystem access, or environment variable usage. All static findings are false positives triggered by pattern matching on documentation text.
2
已扫描文件
1,305
分析行数
3
发现项
5
审计总数
风险因素
📁 文件系统访问 (1)
⚙️ 外部命令 (97)
SKILL.md:68 SKILL.md:68 SKILL.md:68 SKILL.md:79 SKILL.md:79 SKILL.md:79 SKILL.md:90 SKILL.md:90 SKILL.md:90 SKILL.md:101 SKILL.md:101 SKILL.md:101 SKILL.md:112 SKILL.md:112 SKILL.md:112 SKILL.md:120 SKILL.md:131-137 SKILL.md:137-139 SKILL.md:139-149 SKILL.md:149-159 SKILL.md:159-173 SKILL.md:173-191 SKILL.md:191-201 SKILL.md:201-203 SKILL.md:203-227 SKILL.md:227-237 SKILL.md:237-239 SKILL.md:239-264 SKILL.md:264-274 SKILL.md:274-276 SKILL.md:276-303 SKILL.md:303-313 SKILL.md:313-319 SKILL.md:319-344 SKILL.md:344-354 SKILL.md:354-356 SKILL.md:356-366 SKILL.md:366-380 SKILL.md:380-390 SKILL.md:390-400 SKILL.md:400-410 SKILL.md:410-420 SKILL.md:420-430 SKILL.md:430-448 SKILL.md:448-476 SKILL.md:476-482 SKILL.md:482-506 SKILL.md:506-512 SKILL.md:512-515 SKILL.md:515-525 SKILL.md:525-543 SKILL.md:543-549 SKILL.md:549-555 SKILL.md:555-561 SKILL.md:561-573 SKILL.md:573-579 SKILL.md:579-590 SKILL.md:590-596 SKILL.md:596-602 SKILL.md:602-610 SKILL.md:610-655 SKILL.md:655-659 SKILL.md:659-680 SKILL.md:680-690 SKILL.md:690-706 SKILL.md:706-712 SKILL.md:712-724 SKILL.md:724-730 SKILL.md:730-739 SKILL.md:739-765 SKILL.md:765-773 SKILL.md:773-781 SKILL.md:781-837 SKILL.md:837-841 SKILL.md:841-877 SKILL.md:877-881 SKILL.md:881-913 SKILL.md:913-921 SKILL.md:921-936 SKILL.md:936-940 SKILL.md:940-952 SKILL.md:952-956 SKILL.md:956-968 SKILL.md:968-972 SKILL.md:972-984 SKILL.md:984-992 SKILL.md:992-999 SKILL.md:999-1003 SKILL.md:1003-1013 SKILL.md:1013-1017 SKILL.md:1017-1030 SKILL.md:1030-1034 SKILL.md:1034-1052 SKILL.md:1052-1080 SKILL.md:1080-1095 SKILL.md:1095-1099 SKILL.md:1099-1118
审计者: claude 查看审计历史 →
质量评分
38
架构
100
可维护性
87
内容
20
社区
100
安全
78
规范符合性
你能构建什么
Coordinate team development
Orchestrate multiple agents working on backend, frontend, database, and tests in parallel with dependency management
Apply structured workflow
Follow systematic phases from requirements through deployment with built-in quality gates and documentation
Automate quality checks
Run automated security scans, performance analysis, and best practices validation on pull requests
试试这些提示
Start development
Use SPARC methodology to develop feature description Start with researcher mode to gather requirements then architect mode for design
Apply TDD cycle
Run tdd mode to implement feature with 90 percent test coverage Follow Red-Green-Refactor cycle
Coordinate swarm
Initialize hierarchical swarm with number agents Spawn architect coder tester and reviewer agents to work on project in parallel
Review and optimize
Run reviewer mode with security and performance checks on code Then use optimizer mode to address any bottlenecks found
最佳实践
- Always batch all related operations in a single message for parallel execution efficiency
- Use memory integration to share architectural decisions across agent sessions
- Maintain minimum 90 percent test coverage and document all architecture decisions
避免
- Avoid spreading related work across multiple messages as this breaks parallel coordination
- Do not skip the research and architecture phases before implementation
- Avoid using minimal implementation mode in production without proper refactoring
常见问题
Is claude-flow required to use this skill
Yes The skill describes workflows that require claude-flow MCP tools or CLI Install via npm first
What test frameworks does TDD mode support
Jest Mocha and Vitest for unit tests Playwright and Cypress for end-to-end testing
Can SPARC integrate with existing CI/CD pipelines
Yes Documenter mode generates CI/CD ready documentation GitHub integration available via MCP tools
How is my code and data protected
SPARC is a methodology skill with no execution capabilities Code stays in your environment via agent tools
Why are my parallel agents not coordinating properly
Ensure all related operations are in one message Use TodoWrite for task tracking Check memory namespace consistency
How does SPARC compare to traditional development methods
SPARC combines TDD with multi-agent orchestration for 2.8-4.4x speed gains while maintaining 90 percent plus test coverage