PantryPilot 安全扫描报告 — 可信 | ClawSafe

0 /100

PantryPilot

Household replenishment planning skill for mainland China that estimates what is running low, maps menus into restock demand, routes items across Meituan/PDD/Taobao, and outputs cheapest/fastest/lowest-friction restock plans

PantryPilot is a legitimate household replenishment planning skill for Chinese e-commerce platforms with no malicious behavior detected. All functionality is clearly documented, no credential harvesting, no external data exfiltration, and no hidden operations.

技能名称PantryPilot

分析耗时35.5s

引擎pi

✓

可以安装

This skill is safe for deployment. No security action required.

资源类型	声明权限	推断权限	状态	证据
文件系统	`NONE`	`NONE`	—	SKILL.md defines no filesystem access; scripts/publish.sh only modifies temp dir…
网络访问	`READ`	`READ`	✓ 一致	SKILL.md: 'inspect public product pages, grocery pages, activity pages'
命令执行	`NONE`	`NONE`	—	No shell execution in skill behavior; scripts/publish.sh is build-only, not skil…
环境变量	`NONE`	`NONE`	—	No environment variable access in skill implementation
技能调用	`NONE`	`NONE`	—	No cross-skill invocation without user consent documented
剪贴板	`NONE`	`NONE`	—	No clipboard access documented or observed
浏览器	`READ`	`READ`	✓ 一致	SKILL.md: 'inspect public product pages, screenshots' with explicit stop before …
数据库	`NONE`	`NONE`	—	No database access documented or required

目录结构

13 文件 · 34.1 KB · 1107 行

Markdown 9f · 989L Shell 1f · 69L JSON 2f · 45L YAML 1f · 4L

├─ ▾ 📁 agents

│ └─ 📋 openai.yaml YAML 4L · 397 B

├─ ▾ 📁 references

│ ├─ 📝 example-prompts.md Markdown 33L · 1.5 KB

│ ├─ 📝 output-patterns.md Markdown 30L · 659 B

│ ├─ 📝 platform-routing.md Markdown 139L · 4.1 KB

│ ├─ 📝 replenishment-framework.md Markdown 178L · 4.6 KB

│ └─ 📝 test-cases.md Markdown 62L · 1.7 KB

├─ ▾ 📁 scripts

│ └─ 🔧 publish.sh Shell 69L · 2.1 KB

├─ 📝 CHANGELOG.md Markdown 8L · 399 B

├─ 📋 clawhub.json JSON 22L · 598 B

├─ 📋 package.json JSON 23L · 506 B

├─ 📝 README.md Markdown 128L · 3.9 KB

├─ 📝 RELEASE.md Markdown 83L · 2.1 KB

└─ 📝 SKILL.md Markdown 328L · 11.6 KB

安全亮点

✓ Clear documentation of all capabilities in SKILL.md

✓ Explicit safety boundaries defined (no login, no payment, no auto-checkout)

✓ No credential harvesting or sensitive data access

✓ No external data exfiltration or C2 communication

✓ No obfuscated code or base64 payloads

✓ No suspicious file paths accessed (~/.ssh, ~/.aws, .env)

✓ Browser access limited to public pages with explicit stop conditions

✓ Standard CI/CD script for publishing only, not skill execution

✓ MIT license with transparent open-source intent