by 安全扫描
OpenClaw
安全
medium confidenceThe skill's claims, requested resources, and runtime instructions are internally consistent (instruction-only visual QC that produces structured JSON), but the SKILL.md is high-level and leaves important operational details unspecified (how images are processed, where inference runs, and how data is handled).
评估建议
This skill appears coherent for camera-based QC and outputs a clear JSON schema, but it omits operational details you should confirm before installing: (1) Where will images be processed — locally in the agent runtime or sent to an external API? If external, request the endpoint and any required credentials and verify they are declared. (2) Ask the author for the actual model/code used (repo link or container). Without code, you cannot audit inference behavior or data handling. (3) Consider priv...详细分析 ▾
✓ 用途与能力
Name, description, README and SKILL.md all describe the same visual QC use case (camera image inspection for xbot.coffee Lite) and the required outputs. The skill does not request unrelated credentials, binaries, or config paths.
ℹ 指令范围
The SKILL.md specifies detection categories and the exact JSON output schema, but it is high-level and does not instruct how to perform inference (no model, no thresholds, no preprocessing), nor does it state whether images should be processed locally or sent to an external service. That vagueness gives the executing agent broad discretion about where and how to process images, which has privacy and exfiltration implications.
✓ 安装机制
Instruction-only skill with no install steps and no code files. Nothing is written to disk by an installer, which lowers installation risk.
✓ 凭证需求
No environment variables, credentials, or config paths are requested — proportional to the stated purpose. Note: in practice, image processing often requires a model or external API and credentials (not declared here). If the agent uses external vision APIs, additional secrets would be required and should be explicitly declared.
✓ 持久化与权限
Skill is not always-enabled and does not request persistent system-wide privileges. It is user-invocable and can be run autonomously per platform defaults, which is normal for skills. There is no indication it modifies other skills or global config.
安全有层次，运行前请审查代码。
运行时依赖
无特殊依赖
版本
latestv1.0.02026/3/27
- Initial release of xbot-camera-qc-pro (v1.0.0). - Provides professional visual inspection for xbot.coffee Lite, supporting detection of drink volume, screen status, table cleanliness, robotic arm status, equipment cleanliness, environment cleanliness, and pest presence. - Outputs structured English JSON for seamless API automation. - Designed for coffee QC, vision, and automation scenarios.
● 无害
安装命令 点击复制
官方npx clawhub@latest install xbot-camera-qc
镜像加速npx clawhub@latest install xbot-camera-qc --registry https://cn.clawhub-mirror.com
技能文档
Detection Rules
- drink_status: normal / insufficient / overflow / foreign_object / spill_outside
- screen_status: normal / off
- table_clean: normal / dirty
- robot_arm: normal / abnormal
- equipment_dirty: normal / dirty
- environment_clean: normal / messy
- pest: none / found
Output JSON
{ "drink_status": "...", "screen_status": "...", "table_clean": "...", "robot_arm": "...", "equipment_dirty": "...", "environment_clean": "...", "pest": "...", "has_anomaly": true/false, "alert_msg": "..." }数据来源：ClawHub ↗ · 中文优化：龙虾技能库
OpenClaw 技能定制 / 插件定制 / 私有工作流定制
免费技能或插件可能存在安全风险，如需更匹配、更安全的方案，建议联系付费定制