← 返回 Skills 市场
623
总下载
2
收藏
2
当前安装
5
版本数
在 OpenClaw 中安装
/install metric-attribution
功能描述
对指标波动进行综合归因诊断，定位导致指标变化的关键因子、维度和外部事件。当用户需要分析指标波动原因、定位指标变化的根因、做因子拆解归因（如 GMV = UV × 转化率 × 客单价）、或理解指标为什么涨跌时，必须使用此 Skill。 触发场景包括但不限于：用户提到"归因""归因分析""波动分析""波动归因""根因...
安全使用建议
This skill appears internally consistent: it will read a single API key (CAN_API_KEY) from your environment and make network requests to gateway.can.aloudata.com to fetch metric metadata and data for attribution analysis. Before installing, confirm you are comfortable storing the API key in ~/.openclaw/.env and that the key's permissions are appropriately scoped (least privilege). If you expect the skill to correlate external events (news, holidays, competitor data), ask how those external sources will be queried — the skill's domain whitelist only includes the Gateway host, so additional endpoints would need explicit allowance. Finally, review any metric query results for sensitive data exposure and rotate the API key if you stop using the skill.
功能分析
Type: OpenClaw Skill
Name: metric-attribution
Version: 1.0.4
The 'metric-attribution' skill is a legitimate data analysis tool designed to diagnose metric fluctuations using the Aloudata CAN gateway. It provides structured instructions for factor decomposition (using Shapley values) and dimension attribution, including Python code for generating waterfall charts with matplotlib. The requested permissions (env:read and network:outbound to gateway.can.aloudata.com) are strictly aligned with its stated purpose of querying business metrics using a user-provided API key.
能力评估
Purpose & Capability
Name/description (metric attribution) match the declared network calls and the single required env var (CAN_API_KEY). The domain whitelist (gateway.can.aloudata.com) and the API-key header usage are coherent with the stated functionality.
Instruction Scope
SKILL.md narrowly instructs the agent to call the Gateway API and run local calculations; it explicitly requires looking up metric/dimension metadata before querying. One minor note: the doc mentions '结合外部事件检索' (combine external event retrieval) but does not enumerate external endpoints — platform domain whitelist only lists the Gateway host, so either external-event lookups are expected to be manual/knowledge-based or additional endpoints would be needed. Otherwise the instructions do not ask the agent to read unrelated secrets or system files beyond the expected ~/.openclaw/.env for CAN_API_KEY.
Install Mechanism
Instruction-only skill with no install spec or code files, so nothing is written to disk or downloaded during install.
Credentials
Only one required environment variable (CAN_API_KEY) is declared and is appropriate for authenticating to the Gateway API. The skill does not request unrelated credentials or config paths.
Persistence & Privilege
always is false and the skill does not request elevated or persistent system privileges. Autonomous invocation is allowed (platform default) but is not combined with other concerning flags.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install metric-attribution - 安装完成后，直接呼叫该 Skill 的名称或使用
/metric-attribution触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.4
- 新增要求：需在 ~/.openclaw/.env 中配置 CAN_API_KEY 作为必需环境变量，明确标注此项为必填。
- 文档中加入 env_vars 字段说明，提升接入指引清晰度。
- 其余功能与接口说明与上一版本一致，无功能变更。
v1.0.3
- Added required permission and metadata fields (`permissions`, `domain_whitelist`, `author`, `homepage`, etc.).
- Updated the skill `version` from "1.0.3" to "1.0.4".
- No changes to core logic or diagnostic flow; the skill functionality remains the same.
- These changes strengthen compliance, security, and metadata consistency.
v1.0.2
metric-attribution 1.0.2 Changelog
- Added a _meta.json file to the repository.
- No changes to core logic or documentation content.
- This version introduces metadata for improved management and integration.
v1.0.1
metric-attribution v1.0.1
- 修正接口信息：统一所有 Gateway API 规则与 metric-query，删除关于 metric-query-v2 的描述。
- 明确 metric-query API 格式、调用铁律与相关说明，保持与最新接口一致。
- 文档表述细节打磨，优化指引一致性与准确性。
- 其余功能和归因分析流程未作调整。
v1.0.0
metric-attribution v1.0.0
- Initial release of the metric-attribution skill.
- Provides a comprehensive diagnostic workflow for identifying key factors, dimensions, and external events behind metric changes.
- Integrates with Gateway API to search and query metric data; requires strict adherence to API authentication and usage rules.
- Outlines a five-step analysis including change confirmation, factor decomposition, dimension drill-down, external event correlation, and comprehensive reporting.
- Delivers clear guidance on when and how to trigger this skill based on user intent for root cause and attribution analysis.
元数据
常见问题
Aloudata CAN SKILLS - metric-attribution 是什么？
对指标波动进行综合归因诊断，定位导致指标变化的关键因子、维度和外部事件。当用户需要分析指标波动原因、定位指标变化的根因、做因子拆解归因（如 GMV = UV × 转化率 × 客单价）、或理解指标为什么涨跌时，必须使用此 Skill。 触发场景包括但不限于：用户提到"归因""归因分析""波动分析""波动归因""根因... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 623 次。
如何安装 Aloudata CAN SKILLS - metric-attribution？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install metric-attribution」即可一键安装，无需额外配置。
Aloudata CAN SKILLS - metric-attribution 是免费的吗？
是的，Aloudata CAN SKILLS - metric-attribution 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Aloudata CAN SKILLS - metric-attribution 支持哪些平台？
Aloudata CAN SKILLS - metric-attribution 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Aloudata CAN SKILLS - metric-attribution？
由 jackyujun（@jackyujun）开发并维护，当前版本 v1.0.4。
推荐 Skills