← 返回 Skills 市场
Ecommerce Return Intelligence
作者
vx：17605205782
· GitHub ↗
· v1.0.0
· MIT-0
166
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install ecommerce-return-intelligence
功能描述
分析退货原因并区分产品问题、预期错配、物流问题和描述问题。；use for ecommerce, returns, analysis workflows；do not use for 伪造订单数据, 替代客服系统.
使用说明 (SKILL.md)
电商退货洞察器
你是什么
你是“电商退货洞察器”这个独立 Skill，负责：分析退货原因并区分产品问题、预期错配、物流问题和描述问题。
Routing
适合使用的情况
- 分析这批退货原因
- 区分产品问题和预期错配
- 输入通常包含：退货记录 CSV、原因文本、商品信息
- 优先产出：退货概览、原因分类、后续动作
不适合使用的情况
- 不要伪造订单数据
- 不要替代客服系统
- 如果用户想直接执行外部系统写入、发送、删除、发布、变更配置，先明确边界，再只给审阅版内容或 dry-run 方案。
工作规则
- 先把用户提供的信息重组成任务书，再输出结构化结果。
- 缺信息时，优先显式列出“待确认项”，而不是直接编造。
- 默认先给“可审阅草案”，再给“可执行清单”。
- 遇到高风险、隐私、权限或合规问题，必须加上边界说明。
- 如运行环境允许 shell / exec，可使用：
python3 "{baseDir}/scripts/run.py" --input \x3C输入文件> --output \x3C输出文件>
- 如当前环境不能执行脚本，仍要基于
{baseDir}/resources/template.md与{baseDir}/resources/spec.json的结构直接产出文本。
标准输出结构
请尽量按以下结构组织结果：
- 退货概览
- 原因分类
- 高发问题
- 可优化项
- 需复核样本
- 后续动作
本地资源
- 规范文件：
{baseDir}/resources/spec.json - 输出模板：
{baseDir}/resources/template.md - 示例输入输出：
{baseDir}/examples/ - 冒烟测试：
{baseDir}/tests/smoke-test.md
安全边界
- 只做本地分析。
- 默认只读、可审计、可回滚。
- 不执行高风险命令，不隐藏依赖，不伪造事实或结果。
安全使用建议
This skill appears internally consistent and safe for local use. Before running: 1) only run the script on directories/files you intend to analyze (avoid system or secrets directories); 2) if inputs contain PII or credentials, sanitize or omit them first; 3) run in an isolated environment if you want extra assurance (e.g., container or workspace copy); 4) review scripts/run.py if you want to confirm there are no network calls or unexpected writes (the script only reads files and can write the output file you specify). If you intend to automate actions (writes/notifications/changes) based on its output, treat the skill's output as a human-reviewed draft and implement explicit safeguards before performing external writes.
功能分析
Type: OpenClaw Skill
Name: ecommerce-return-intelligence
Version: 1.0.0
The skill bundle is a legitimate tool designed for analyzing ecommerce return data and auditing file structures. The primary script, `scripts/run.py`, performs local file processing, CSV parsing, and includes a defensive pattern-matching feature to identify potential risks (like hardcoded secrets or dangerous shell commands) within input data. There is no evidence of data exfiltration, network communication, or malicious prompt injection; the code is well-structured and adheres to its stated purpose.
能力评估
Purpose & Capability
Name/description (returns analysis) aligns with the included assets: resources/spec.json, template.md, examples, and scripts/run.py. The only declared binary is python3 which is necessary to run the helper script; there are no unrelated environment variables or credentials requested.
Instruction Scope
SKILL.md confines work to local analysis, explicitly advises read-only/auditable outputs, and instructs running the bundled python script or using the templates if execution is not allowed. The script reads files (including CSV/TSV/MD/py/js/sh), recursively lists files in directories, and performs pattern scanning — it does not execute discovered scripts or reach out to network endpoints. Because the script will read any file/directory the user supplies, users should avoid pointing it at system root or directories containing secrets/credentials.
Install Mechanism
There is no install spec and no network downloads; the skill is instruction+script-only and relies on the Python standard library. No archives or external package installs are performed.
Credentials
The skill requires no environment variables or external credentials. All operations are local and driven by user-provided input files; no secret-sounding env names or unrelated cloud credentials are requested.
Persistence & Privilege
always is false and the skill does not attempt to modify other skills or global agent configuration. The script can write an output file when given --output, which is expected behavior for a reporting tool.
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 ecommerce-return-intelligence - 安装完成后，直接呼叫该 Skill 的名称或使用
/ecommerce-return-intelligence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
v1.0.0
Initial release of ecommerce-return-intelligence.
- Provides analysis of e-commerce return reasons, categorizing issues as product problem, expectation mismatch, logistics, or description issue.
- Outputs structured reports with overview, cause classification, frequent problems, suggested improvements, samples requiring review, and recommended actions.
- Operates in a read-only, auditable, and safe manner; explicitly avoids fabricating data or replacing customer service systems.
- Utilizes Python script if possible, otherwise generates results using provided templates and specifications.
- Clearly lists missing information for user confirmation and always provides reviewable drafts before executable recommendations.
元数据
常见问题
Ecommerce Return Intelligence 是什么？
分析退货原因并区分产品问题、预期错配、物流问题和描述问题。；use for ecommerce, returns, analysis workflows；do not use for 伪造订单数据, 替代客服系统. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 166 次。
如何安装 Ecommerce Return Intelligence？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install ecommerce-return-intelligence」即可一键安装，无需额外配置。
Ecommerce Return Intelligence 是免费的吗？
是的，Ecommerce Return Intelligence 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Ecommerce Return Intelligence 支持哪些平台？
Ecommerce Return Intelligence 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（darwin, linux, win32）。
谁开发了 Ecommerce Return Intelligence？
由 vx：17605205782（@52yuanchangxing）开发并维护，当前版本 v1.0.0。
推荐 Skills