← Back to Skills Marketplace
wteng2286

eric-compliance-suite

by wteng2286 · GitHub ↗ · v1.0.1 · MIT-0
cross-platform ⚠ suspicious
231
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install eric-compliance-suite
Description
睿观(ERiC) 全功能合规检测套件。集成外观专利(D001)、发明专利(I001)、图形商标(L001)、文本商标+替换词(T001/T002)、版权(C001)、政策合规(P001-P007) 六大检测能力。当用户需要进行任何知识产权合规检测(专利、商标、版权)或电商平台政策合规审查时触发此 skill。
README (SKILL.md)

睿观 ERiC 合规检测套件

API Key 获取方式

关注微信公众号【睿观AI】,发送消息:"睿观Skill",即可获取 API Key。

获取后请设置环境变量:

export ERIC_API_TOKEN="你获取到的 API Key"

扣点说明(首次使用必读)

检测模式 基础扣点 开启雷达后 说明
D001 外观专利 10 点 15 点 默认开启雷达
I001 发明专利 10 点 - 无雷达功能
L001 图形商标 10 点 15 点 默认开启雷达
T001 文本商标 1 点 - 无雷达功能
T002 商标替换词 1 点 - 按词计费
C001 版权检测 1 点 2 点 默认开启雷达
P001 政策-纯图 1 点 - -
P002 政策-纯文本 5 + 特征词数×2 点 - 图文同时入参仍为 5+2n 点

重要提示:

  • 带雷达的检测模式(D001、L001、C001)默认开启雷达以获得更精准的结果
  • 检测开始前会提示预扣点数,检测完成后会显示实际扣点
  • 如客户端超时导致重试,会说明实际扣费情况

功能导航

根据用户需求选择对应的 API:

用户需求 API 输入类型 参考文档
产品图片查外观专利侵权 D001 图片 design-patent.md
产品标题/描述查发明专利侵权 I001 文本 invention-patent.md
产品图片查图形商标(logo)侵权 L001 图片 logo-detection.md
产品标题/描述查文本商标风险 T001 文本 trademark-detection.md
为高风险商标词找安全替换词 T002 文本 trademark-detection.md
产品图片查版权画侵权 C001 图片 copyright-detection.md
产品图片查违禁品(枪械配件等) P001 图片 policy-detection.md
产品文本查平台销售政策合规 P002 文本 policy-detection.md
管理自定义风险特征词 P004-P007 文本 policy-detection.md

上下文保护

  • 禁止将 base64 图片、大体积二进制数据读入 Agent 上下文。 图片编码和 API 调用必须在代码执行环境中完成。
  • API 响应过大时,先在代码执行环境中提取关键字段摘要再呈现给用户。
  • 涉及图片的接口:D001、L001、C001、P001。

认证配置

所有接口共用同一 Token。调用前需确认环境变量 ERIC_API_TOKEN 已设置。若未设置,提示用户:

未检测到 ERIC_API_TOKEN 环境变量。请先完成配置:
1. 登录睿观平台 https://eric-bot.com 获取 API Token
2. 设置环境变量:export ERIC_API_TOKEN=your_token

所有请求的公共 Headers:Content-Type: application/jsonToken: \x3CAPI_TOKEN>

调用方式

统一入口脚本 scripts/detect.py,通过子命令区分 API:

# D001 外观专利检测
python scripts/detect.py d001 /path/to/image.png --regions US GB --top 50

# I001 发明专利检测
python scripts/detect.py i001 --title "产品标题" --description "产品描述" --regions US

# L001 图形商标检测
python scripts/detect.py l001 /path/to/image.png --top 20 --regions US --enable-radar

# T001 文本商标检测 (--auto-safe-words 自动为高风险词调 T002)
python scripts/detect.py t001 --title "Ps4 Wireless Controller" --text "描述..." --regions US JP --auto-safe-words

# T002 商标替换词
python scripts/detect.py t002 --title "产品标题" --text "产品描述" --trademark "商标词"

# C001 版权检测
python scripts/detect.py c001 /path/to/image.png --top 100 --enable-radar

# P001 政策合规-纯图检测
python scripts/detect.py p001 /path/to/image.png

# P002 政策合规-纯文本检测
python scripts/detect.py p002 --title "产品标题" --description "描述" --sites us jp

# P004 风险特征词联想
python scripts/detect.py p004 "模糊词"

# P005/P006/P007 特征词管理
python scripts/detect.py p005 "特征词"
python scripts/detect.py p006 123
python scripts/detect.py p007 --per-page 50 --page 1

所有子命令均支持 --json 输出原始 API 响应。使用 python scripts/detect.py \x3C子命令> --help 查看完整参数。

D001 外观专利检测

通过产品图片搜索相似外观专利,按相似度排序,单次最多 500 条。

  • URL: POST https://saas.eric-bot.com/v1.0/eric-api/patent/design/v1/detection
  • 必需参数: img_64lis(图片base64数组), regions(国家代码), top_number(1-500), enable_tro(bool), query_mode("hybrid"/"physical"/"line")
  • 可选参数: product_title, product_description, top_loc, patent_status, source_language, enable_radar
  • 关键响应: data.list[]similarity, patent_prod, publication_number, patent_validity, tro_holder, radar_result
  • 风险判断: similarity > 0.8radar_result.same = truetro_holder = true → 高风险
  • 费用: 10 点/次(默认开启雷达 15 点/次
  • 默认站点: US

详细参数、响应字段、错误码和代码示例见 references/design-patent.md

I001 发明专利检测

通过产品标题和描述搜索相似发明专利,当前仅支持 US。

  • URL: POST https://saas.eric-bot.com/v1.0/eric-api/patent/utility/v1/detection
  • 必需参数: product_title, product_description, regions(仅["US"]), top_number(1-500)
  • 关键响应: data.data[]similarity(number), title, abstract, publication_number, cpc_classification[], specification_url, claims_url
  • 风险判断: similarity > 0.8 → 高风险
  • 费用: 10 点/次
  • 默认站点: US(当前仅支持 US)

详细参数、响应字段、错误码和代码示例见 references/invention-patent.md

L001 图形商标检测

通过产品图片识别 logo 区域并搜索相似已注册图形商标,单次最多 100 条。

  • URL: POST https://saas.eric-bot.com/v1.0/eric-api/trademark/graphic/v1/detection
  • 必需参数: base64_image, top_number(1-100)
  • 可选参数: product_title, trademark_name, regions(15个国家), enable_localizing, enable_radar
  • 关键响应: data.detection_results[]top_graphic_trademarks[]graphic_trademarks[]similarity, trademark_name, applicant_name, trade_mark_status, nice_class[], sub_radar_result
  • 风险判断: sub_radar_result = "high_risk"similarity > 0.8 → 高风险
  • 费用: 10 点/次(默认开启雷达 15 点/次
  • 默认站点: 全部(US, WO, ES, GB, DE, IT, CA, MX, EM, AU, FR, JP, TR, BX, CN)

详细参数、响应字段、错误码和代码示例见 references/logo-detection.md

T001 文本商标检测 + T002 替换词

T001 检测产品文本中的商标词及风险等级;T002 为高风险词生成安全替换词。

T001 文本商标检测

  • URL: POST https://saas.eric-bot.com/v1.0/eric-api/trademark/text/v1/detection
  • 必需参数: product_title(≤300字符)
  • 可选参数: product_text(≤5000字符), regions(支持15个国家: AU, BX, CA, DE, EM, ES, FR, GB, IT, JP, MX, TR, US, WO, CN)
  • 关键响应: data.text_trademarks[]trademark_name, highest_mode_score(0-5), status, is_active_holder, is_famous, region_score[]data.text_trademark_radar(0=低风险/1=待核查/2=高风险)
  • 费用: 1 点/次
  • 超时: 90 秒
  • 默认站点: US

T002 商标替换词

  • URL: POST https://saas.eric-bot.com/v1.0/eric-api/trademark/text/v1/safe-words-generation
  • 必需参数: product_title, product_text, trademark_name(T001检出的单个商标词)
  • 关键响应: data.words[] 推荐替换词
  • 费用: 1 点/次

工作流: 首次检测只调用 T001(扣 1 点),不自动调用 T002。如需获取替换词,用户需手动指定 --auto-safe-words 参数,此时会为 highest_mode_score >= 3 的词调 T002 获取替换建议。

详细参数、响应字段、错误码和代码示例见 references/trademark-detection.md

C001 版权检测

通过产品图片搜索相似版权画作,单次最多 200 条。

  • URL: POST https://saas.eric-bot.com/v1.0/eric-api/copyright/v1/detection
  • 必需参数: img_64lis(图片base64数组), top_number(默认100, 最大200), enable_radar(bool)
  • 关键响应: data.list[]similarity(相似度), path(版权画图片), rights_owner, copyright_code, copyright_url, sub_radar_result, tro_holder
  • 风险判断: similarity > 0.8sub_radar_result = "high_risk" → 高风险
  • 费用: 1 点/次(默认开启雷达 2 点/次

详细参数、响应字段、错误码和代码示例见 references/copyright-detection.md

P001-P007 政策合规检测

检测产品是否违反电商平台销售政策,并管理风险特征词。

P001 纯图检测

  • URL: POST https://saas.eric-bot.com/v1.0/eric-api/policy-compliance/v1/gun-parts-search
  • 必需参数: base64_image, type(当前仅 ["gun_parts"])
  • 关键响应: data.list[]pd_title, pd_img_oss_url, cosine
  • 后续: cosine >= 0.4 时建议调 P002 确认具体违反的政策
  • 费用: 1 点/次

P002 纯文本检测

  • URL: POST https://saas.eric-bot.com/v1.0/eric-api/policy-compliance/v1/detection
  • 必需参数: product_title, platform_sites(如 {"amazon":["us","jp"]}), feature_detect(含 enable, features)
  • 可选参数: product_description, type(P001有结果时传), product_title_suspected
  • 关键响应: data.list[]prohibited(1=禁售), compliance(1=限售), reason, content_url, name
  • 费用: 5 + 特征词数×2 点/次(图文同时入参检测仍为 5+2n 点,不分开计费)
  • 默认站点: us
  • 支持站点: br, fr, au, us, uk(非gb), jp, it, es, mx, de, ca

P004-P007 风险特征词管理

操作 URL 路径 参数
P004 联想 policy-compliance/feature/v1/suggestion word(模糊词)
P005 保存 policy-compliance/feature/v1/save word
P006 删除 policy-compliance/feature/v1/delete id
P007 列表 policy-compliance/feature/v1/list per_page, page

URL 前缀: https://saas.eric-bot.com/v1.0/eric-api/

详细参数、响应字段、错误码和代码示例见 references/policy-detection.md

通用错误处理

所有接口共用以下错误码:

code 含义 处理建议
30007 Token 无效 检查 API Token
100007 参数错误 检查必填字段
4000011-4000014 积分不足/扣点失败 登录 eric-bot.com 充值
4000020-4000021 图片上传失败 检查图片格式和大小
5000504 请求超时 稍后重试
0 / 500 系统异常 稍后重试

各接口专属错误码见对应 references/ 文档。

Usage Guidance
This skill appears to do what it claims (IP and policy checks) and communicates which external endpoints it calls, but exercise caution before installing or running it: - The SKILL.md and scripts require ERIC_API_TOKEN, yet the registry metadata did not declare any required env vars — ensure you are comfortable providing an API token and verify the token's issuer and scope first. - The script uploads images and product text to saas.eric-bot.com. Do not send sensitive or proprietary images/text unless you trust that external service and its privacy/retention policies. - The script will auto-install the Python 'requests' package via pip on first run. If you prefer not to allow runtime package installs, run the script in a controlled environment or pre-install dependencies yourself. - Source/homepage are missing and the owner is unknown; prefer skills with verifiable publishers or published project pages. If you need higher assurance, inspect scripts/detect.py locally, run it in a sandboxed container, or request the developer/publisher identity and privacy policy before use.
Capability Analysis
Type: OpenClaw Skill Name: eric-compliance-suite Version: 1.0.1 The eric-compliance-suite is a legitimate tool designed for intellectual property (patents, trademarks, copyrights) and e-commerce policy compliance checking. The main script, scripts/detect.py, facilitates communication with the ERiC API (saas.eric-bot.com) and includes standard functionality for processing images and text. While the script contains a routine to auto-install the 'requests' library via pip if it is missing, this behavior is a common utility pattern in this context and lacks malicious intent. The instructions in SKILL.md are focused on operational efficiency, such as managing agent context and summarizing large API responses, and do not contain harmful prompt injections.
Capability Assessment
Purpose & Capability
Name, description, SKILL.md, references, and scripts all consistently implement patent/trademark/copyright/policy checks against the same saas.eric-bot.com API endpoints — functionality is coherent with the stated purpose. However the registry metadata claims 'Required env vars: none' while the runtime docs and script require ERIC_API_TOKEN.
Instruction Scope
SKILL.md and scripts are explicit about what data is sent and which endpoints are used (design/trademark/copyright/policy endpoints). The runtime instructions avoid embedding raw base64 in the agent context and direct the code to handle images locally or by URL. There is no broad, vague instruction to read arbitrary system files or other credentials.
Install Mechanism
There is no formal install spec (instruction-only), which lowers install risk. The included scripts auto-install the Python 'requests' package via pip at runtime (subprocess.run), which will modify the environment. That behavior is common but means the skill will execute package installs when first run.
Credentials
The skill actually requires a secret API token (ERIC_API_TOKEN) to call external APIs, but the registry metadata lists no required env vars or primary credential — an important mismatch. The single token requested is proportional to the service's purpose, but the omission in declared requirements and the fact the skill will upload image content and product text to an external service are privacy/credential considerations the user should be aware of.
Persistence & Privilege
The skill is not set 'always: true' and does not request elevated platform privileges. It does not persistently modify other skills or system-wide settings. The only notable side-effect is installing a Python package (requests) when executed.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install eric-compliance-suite
  3. After installation, invoke the skill by name or use /eric-compliance-suite
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
fix: handle radar_result=None to avoid AttributeError in D001
v1.0.0
eric-compliance-suite v1.0.0 - 首次发布:集成专利、商标、版权、电商政策合规等六大检测能力
Metadata
Slug eric-compliance-suite
Version 1.0.1
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 2
Frequently Asked Questions

What is eric-compliance-suite?

睿观(ERiC) 全功能合规检测套件。集成外观专利(D001)、发明专利(I001)、图形商标(L001)、文本商标+替换词(T001/T002)、版权(C001)、政策合规(P001-P007) 六大检测能力。当用户需要进行任何知识产权合规检测(专利、商标、版权)或电商平台政策合规审查时触发此 skill。 It is an AI Agent Skill for Claude Code / OpenClaw, with 231 downloads so far.

How do I install eric-compliance-suite?

Run "/install eric-compliance-suite" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is eric-compliance-suite free?

Yes, eric-compliance-suite is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does eric-compliance-suite support?

eric-compliance-suite is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created eric-compliance-suite?

It is built and maintained by wteng2286 (@wteng2286); the current version is v1.0.1.

💬 Comments