← Back to Skills Marketplace
155
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install label-compliance-audit
Description
智能审核食品包装标签的合规性，支持图片/PDF设计稿；当用户需要审核标签强制性标识、营养标签规范性、声称用语合规性或格式要求时使用
README (SKILL.md)
标签合规审核
任务目标
- 本 Skill 用于：食品包装标签的智能合规审核
- 能力包含：识别标签内容、对照GB7718/GB28050标准检查、生成审核报告
- 触发条件：用户上传包装设计稿并要求审核标签合规性
前置准备
- 依赖说明：无需外部依赖，使用智能体原生能力
- 非标准文件/文件夹准备：无需额外准备
操作步骤
-
标准流程：
- 读取设计稿
- 使用图像识别工具提取标签中的所有文字和图片内容
- 识别关键信息：品名、配料表、净含量、生产者信息、营养标签、声称用语等
- 选择审核模式
- 模式A：智能体自主审核
- 基于以下参考标准进行全面审核
- 生成结构化审核报告
- 模式B：逐项清单比对
- 参考
assets/audit-checklist.md中的审核清单 - 逐项对照并记录审核结果
- 参考
- 模式A：智能体自主审核
- 执行审核（参考以下文档）
- 强制性标识完整性：见 references/gb7718-standard.md
- 营养标签规范性：见 references/gb28050-standard.md
- 声称用语合规性：见 references/claims-guideline.md
- 格式要求（字体大小、对比度）：见 references/gb7718-standard.md
- 过敏原标识：见 references/gb7718-standard.md
- 生成审核报告
- 列出所有不符合项
- 提供整改建议
- 标注风险等级（严重/一般/轻微）
- 读取设计稿
-
可选分支：
- 当设计稿为PDF时：先提取页面内容，再进行图像识别
- 当需要详细NRV计算：参考GB28050标准中的计算公式逐项验证
资源索引
- 领域参考：
- references/gb7718-standard.md（何时读取：审核强制性标识、格式要求、过敏原）
- references/gb28050-standard.md（何时读取：审核营养标签、NRV计算）
- references/claims-guideline.md（何时读取：审核声称用语）
- 资产模板：
- assets/audit-checklist.md（用途：逐项清单比对审核）
注意事项
- 审核时需全面识别标签中的所有文字信息，包括小字说明
- 营养标签的NRV%需根据GB28050中的计算方法逐项验证
- 声称用语必须满足特定条件才能使用（如"无糖"需≤0.5g/100g）
- 格式要求需关注字体高度（≥1.8mm）和背景对比度
- 过敏原标识需明确标注，采用加粗或特殊字体
使用示例
示例1：智能体自主审核
- 功能说明：上传设计稿，智能体自动完成全面审核
- 执行方式：智能体读取参考标准，自主判断并生成报告
- 关键要点：
- 全面识别标签内容
- 逐项对照标准要求
- 生成结构化审核报告
- 示例：用户上传饼干包装设计稿，要求审核标签合规性
示例2：逐项清单比对
- 功能说明：使用审核清单逐项核对标签内容
- 执行方式：参考 assets/audit-checklist.md 逐项检查
- 关键要点：
- 按清单顺序逐项确认
- 记录每一项的审核结果
- 汇总不符合项
- 示例：用户要求使用清单方式审核饮料标签
示例3：营养标签专项审核
- 功能说明：重点审核营养标签的规范性和NRV计算
- 执行方式：重点读取 references/gb28050-standard.md
- 关键要点：
- 验证能量、蛋白质、脂肪等核心营养素标注
- 计算并核实NRV%
- 检查营养声称的合规性
- 示例：用户需要验证酸奶营养标签的NRV%计算准确性
Usage Guidance
This skill is internally consistent and contains helpful reference documents and a checklist. Before installing or using it, verify where image/PDF processing (OCR/vision) runs: if the agent will call a third‑party OCR service, avoid uploading sensitive images or confirm the service's privacy policy. Test with non-confidential samples to confirm output quality and check that generated NRV calculations and claims checks match your local/regional regulations. If you need guarantees about where data is processed, ask the vendor or platform how OCR is implemented (local vs external).
Capability Analysis
Type: OpenClaw Skill
Name: label-compliance-audit
Version: 1.0.1
The skill bundle is a legitimate tool designed for auditing food packaging labels against Chinese national standards (GB7718 and GB28050). The instructions in SKILL.md and the reference documents in the references/ directory are strictly focused on label compliance, OCR text extraction, and report generation, with no evidence of malicious intent, data exfiltration, or harmful prompt injection.
Capability Assessment
Purpose & Capability
Name/description match the included reference documents and checklist (GB7718, GB28050, claims guideline). The skill does not request unrelated credentials, binaries, or config paths — all requested capabilities are appropriate for a label-audit skill.
Instruction Scope
SKILL.md stays within the stated purpose (extract text/images, compare with standards, produce a report). It references using an image-recognition tool but does not specify which tool or whether processing occurs locally or via a third party; that is a usability/privacy note rather than an incoherence.
Install Mechanism
Instruction-only skill with no install spec and no code files—nothing is written to disk or downloaded during install, which is proportionate for a policy/checklist-based auditor.
Credentials
No environment variables, credentials, or config paths are required. The skill's resource usage (reading included reference docs and user-uploaded designs) is proportionate to its purpose.
Persistence & Privilege
always is false and the skill does not request persistent system-wide privileges or modify other skills. Autonomous invocation is allowed by platform default but is not combined with any elevated access here.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install label-compliance-audit - After installation, invoke the skill by name or use
/label-compliance-audit - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
Version 1.0.1 of label-compliance-audit
- No file changes detected in this release.
- Documentation and core workflow remain unchanged.
- Maintains all previous features for intelligent food label compliance auditing.
v1.0.0
label-compliance-audit v1.0.0
- Initial release: Smart compliance auditing for food packaging labels, supporting image/PDF files.
- Checks mandatory elements, nutrition label accuracy, claims wording, format, and allergen identification based on GB7718/GB28050 standards.
- Offers two audit modes: automatic full review or checklist-based item-by-item comparison.
- Generates structured audit reports highlighting nonconformities, recommendations, and risk levels.
- No external dependencies required.
Metadata
Frequently Asked Questions
What is label audit食品标签审核?
智能审核食品包装标签的合规性，支持图片/PDF设计稿；当用户需要审核标签强制性标识、营养标签规范性、声称用语合规性或格式要求时使用. It is an AI Agent Skill for Claude Code / OpenClaw, with 155 downloads so far.
How do I install label audit食品标签审核?
Run "/install label-compliance-audit" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is label audit食品标签审核 free?
Yes, label audit食品标签审核 is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does label audit食品标签审核 support?
label audit食品标签审核 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created label audit食品标签审核?
It is built and maintained by Tinker (@cloudyxuq); the current version is v1.0.1.
More Skills