validate-plugin
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseValidate Plugin
验证插件
Validate that a plugin follows the correct Claude Code plugin format.
验证插件是否遵循正确的Claude Code插件格式。
When to use
使用场景
After creating or modifying a plugin, run validation to catch structural issues before publishing.
在创建或修改插件后,运行验证以在发布前发现结构问题。
Checks performed
执行的检查项
- Directory structure — exists at plugin root
.claude-plugin/plugin.json - plugin.json schema — required fields present (name, description, version)
- Skills — each skill in has a matching
plugin.json.skillsskills/<name>/SKILL.md - Commands — each command in has a matching
plugin.json.commandscommands/<name>.md - Agents — each agent in has a matching
plugin.json.agentsagents/<name>.md - SKILL.md frontmatter — each skill has ,
name, anddescriptionallowed-tools - Agent frontmatter — each agent has ,
name, anddescriptionmodel - No files in wrong locations — skills/commands/agents not inside
.claude-plugin/ - MCP tool references — tools in are valid
allowed-toolsidentifiersmcp__claude-flow__*
- 目录结构 — 插件根目录下存在文件
.claude-plugin/plugin.json - plugin.json schema — 包含必填字段(name、description、version)
- Skills — 中的每个skill都有对应的
plugin.json.skills文件skills/<name>/SKILL.md - Commands — 中的每个command都有对应的
plugin.json.commands文件commands/<name>.md - Agents — 中的每个agent都有对应的
plugin.json.agents文件agents/<name>.md - SKILL.md 前置元数据 — 每个skill都包含、
name和description字段allowed-tools - Agent 前置元数据 — 每个agent都包含、
name和description字段model - 无文件放置错误 — skills/commands/agents文件未存放在目录内
.claude-plugin/ - MCP工具引用 — 中的工具是有效的
allowed-tools标识符mcp__claude-flow__*
Steps
步骤
- Read the plugin's
plugin.json - For each skill, command, and agent declared, verify the file exists
- For each SKILL.md, verify frontmatter has required fields
- For each agent .md, verify frontmatter has required fields
- Report pass/fail for each check with actionable fix suggestions
- 读取插件的文件
plugin.json - 针对每个声明的skill、command和agent,验证对应文件是否存在
- 针对每个SKILL.md文件,验证前置元数据包含必填字段
- 针对每个agent的.md文件,验证前置元数据包含必填字段
- 报告每项检查的通过/失败结果,并提供可操作的修复建议