From omni
Orchestrates registered AI-friendly architecture metric skills to evaluate code repositories across dimensions in modes like full, default, or custom sets, aggregating into unified JSON reports. Use for comprehensive arch quality assessments or CI pipelines.
npx claudepluginhub zte-aicloud/co-omnispec --plugin omniThis skill uses the workspace's default tool permissions.
以代码库为粒度，编排所有已注册的 AI 友好架构度量 skill，输出跨维度聚合报告。
Executes AI-friendly architecture metrics on code repositories, generating aggregated multi-dimensional reports with scores, grades, and violation summaries. Useful for codebase architecture audits.
Orchestrates extensible code quality audits: discovers dimensions, builds DAG for phased parallel execution via subagents, each in isolated context window.
Conducts full codebase health audit across architecture, security, code quality, dependencies, test coverage. Produces scored report with letter grades and prioritized remediation. Use for existing codebases or before releases.
Share bugs, ideas, or general feedback.
以代码库为粒度，编排所有已注册的 AI 友好架构度量 skill，输出跨维度聚合报告。
| 参数 | 类型 | 必填 | 说明 |
|---|---|---|---|
project_path | string | ✅ | 待分析项目根目录（绝对路径） |
execute_mode | enum | ❌ | 执行模式：default（默认）/ all / dimension / skills |
dimension | string | ❌ | 指定维度（execute_mode=dimension 时必填） |
skill_ids | string | ❌ | 逗号分隔的 skill ID（execute_mode=skills 时必填） |
output_path | string | ❌ | 最终报告路径（默认 output/arch-measure-report.json） |
| Step | 职责 | 执行者 | 文档 | 输入 | 输出 |
|---|---|---|---|---|---|
| 01 | 解析待执行 skill 列表 | 脚本 scripts/resolve-skills.py | step01 | config/metric-registry.json + 执行参数 | state/resolved-skills.json |
| 02 | 顺序执行各度量 skill | Main Agent | step02 | state/resolved-skills.json | output/<skill>/summary.json |
| 03 | 聚合所有结果 | 脚本 scripts/aggregate-metrics.py | step03 | state/resolved-skills.json + skill 输出文件 | output/arch-measure-report.json |
执行 scripts/resolve-skills.py，根据执行参数从注册表解析 skill 列表：
python scripts/resolve-skills.py \
--registry config/metric-registry.json \
--output state/resolved-skills.json \
[--all | --dimension <维度名> | --skills <skill_id,...>]
验证 state/resolved-skills.json 存在且可解析后，进入 Step 2。
读取 state/resolved-skills.json 中的 resolved 列表，按顺序调用每个 skill：
对 resolved 列表中的每个 skill：
调用技能 `<skill_id>`
参数：
project_path: <project_path>
output_path: <output_path_hint>
等待完成，验证输出文件存在
若失败：记录到失败列表，继续下一个 skill
容错策略：单个 skill 失败不阻断其余 skill 执行。
执行 scripts/aggregate-metrics.py，读取各 skill 输出并生成总报告：
python scripts/aggregate-metrics.py \
--resolved-skills state/resolved-skills.json \
--output-dir output \
--output output/arch-measure-report.json \
--project-path <project_path>
验证 output/arch-measure-report.json 存在且可解析。
新增度量 skill 只需在 config/metric-registry.json 追加记录，无需修改本文件：
{
"skill_id": "ai-friendly-metric-xxx",
"display_name": "新指标名称",
"dimension": "目标维度",
"tags": ["default"],
"enabled": true,
"description": "指标说明",
"output_path_hint": "output/xxx/summary.json"
}
state/
resolved-skills.json # Step 1 产物
output/
srp/
summary.json # SRP skill 产物（aia_metric_fact 格式）
arch-measure-report.json # 最终聚合报告（aia_component_summary 格式）
output/arch-measure-report.json）采用双 key 顶层结构，字段定义详见 config/data_model.md：
| Key | 类型 | 说明 |
|---|---|---|
aia_metric_fact | list | 各 skill 完整详情（identity_info、execution_ctx、core_metrics、evaluation_details、violation_records、scan_statistics）；失败/跳过的 skill 有最小占位 fact |
aia_component_summary | dict | 组件汇总（identity_info、scan_result、dimension_data、relation_mapping、_meta）；各字段从 aia_metric_fact 列表推导聚合；_meta 为本 skill 私有扩展字段 |
| 场景 | 处理 |
|---|---|
| Step 1 失败（registry 解析错误） | 阻断流程，报告错误 |
| 某 skill 执行失败 | 容错继续，记录 _meta.failed_skills |
| 所有 skill 均失败 | _meta.execute_status: "failed" |
| Step 3 失败 | 各 skill 原始输出仍存在，可手动查阅 |