From omni
Measures AI-friendly architecture across multiple dimensions (e.g., responsibility, coupling) and generates an aggregated report with scores, grades, and violation counts. Useful for evaluating codebase structure and identifying improvement areas.
How this skill is triggered — by the user, by Claude, or both
Slash command
/omni:arch-measureThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
对代码库执行 AI 友好架构全面度量,生成跨维度聚合报告。
对代码库执行 AI 友好架构全面度量,生成跨维度聚合报告。
/arch-measure [选项]
| 参数 | 说明 | 示例 |
|---|---|---|
--project-path <路径> | 待分析项目根目录(绝对路径)。未指定时使用当前工作目录 | --project-path /path/to/repo |
--all | 执行所有 enabled: true 的度量 skill | --all |
--dimension <维度名> | 仅执行指定维度下的度量 skill | --dimension 职责维度 |
--skills <id,...> | 强制执行指定 skill(逗号分隔,忽略 enabled 状态) | --skills ai-friendly-component-srp-orchestrate |
| 无参数 | 执行默认集(tags 包含 "default" 且 enabled: true) | /arch-measure |
project_path(若未提供,询问用户或使用当前目录)--all → 全量模式--dimension <维度名> → 按维度模式--skills <id,...> → 按 skill 模式解析参数后,调用技能 ai-friendly-arch-measure,传入:
调用技能 `ai-friendly-arch-measure`
参数:
project_path: <project_path>
execute_mode: <default|all|dimension|skills>
dimension: <维度名>(按维度模式时)
skill_ids: <id,...>(按 skill 模式时)
output_path: output/arch-measure-report.json
skill 执行完成后,读取 output/arch-measure-report.json,展示以下摘要:
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
AI 友好架构度量报告
项目:<component_name> 路径:<component_repo>
执行模式:<execute_mode> 评级:<overall_grade>
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
综合得分:<total_score_avg>
总违规数:<total_violations>(P0: <p0_total>,P1: <p1_total>)
维度得分:
职责维度 <score> [<status>]
...
跳过 skill:<skipped_skills>(如有)
失败 skill:<failed_skills>(如有)
详细报告:output/arch-measure-report.json
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
| 等级 | 分数区间 | 含义 |
|---|---|---|
| S | ≥ 0.9 | 优秀 |
| A | ≥ 0.75 | 良好 |
| B | ≥ 0.6 | 一般 |
| C | ≥ 0.4 | 待改进 |
| D | < 0.4 | 严重问题 |
# 默认集(执行已启用的默认 skill)
/arch-measure --project-path /path/to/repo
# 全量执行所有已启用 skill
/arch-measure --project-path /path/to/repo --all
# 仅执行职责维度相关 skill
/arch-measure --project-path /path/to/repo --dimension 职责维度
# 指定 skill 执行
/arch-measure --project-path /path/to/repo --skills ai-friendly-component-srp-orchestrate
npx claudepluginhub zte-aicloud/co-omnispec --plugin omniOrchestrates multi-dimensional architecture quality assessment across registered metric skills. Supports full, dimension, skill, and default execution modes with aggregated reporting.
Orchestrates extensible code quality audits: discovers dimensions, builds DAG for phased parallel execution via subagents, each in isolated context window.
Analyzes codebase architecture via multi-agent specialists on structure, coupling, integration, error handling, security; verifies findings, reports strengths and flaws with evidence.