feishu-doc-block-writer 安全扫描报告 — 低风险 | ClawSafe

20 /100

feishu-doc-block-writer

飞书文档 Block 拆分写入技能 - 自动将长内容拆分为多个 Blocks 写入飞书文档

Documentation-only skill referencing non-existent implementation scripts - no actual malicious code present, but SKILL.md promises functionality that doesn't exist.

技能名称feishu-doc-block-writer

分析耗时35.8s

引擎pi

✓

可以安装

This skill references scripts/block-writer.py that are not included in the package. Either add the implementation files or update SKILL.md to clarify this is a documentation-only reference skill.

安全发现 2 项

严重性	安全发现	位置
低危	Missing Implementation Files 文档欺骗 SKILL.md references a script at scripts/block-writer.py and shows example bash commands, but no script files exist in the skill package. `python block-writer.py --title "文档标题" --content "长文本内容..."` → Add the implementation script or remove the script references from documentation.	`SKILL.md:53`
提示	Unverified Tool Usage 文档欺骗 SKILL.md mentions using feishu_doc tool with create/append methods but skill.json doesn't declare allowed-tools mapping. `使用 create 创建空文档，使用 append 逐块追加内容` → Add allowed-tools declaration to skill.json if this skill invokes other skills/tools.	`SKILL.md:13`

资源类型	声明权限	推断权限	状态	证据
文件系统	`NONE`	`NONE`	—	SKILL.md mentions script paths but no files exist
网络访问	`NONE`	`NONE`	—	No network calls in documentation or code
命令执行	`NONE`	`NONE`	—	SKILL.md shows bash examples but no scripts exist
环境变量	`NONE`	`NONE`	—	No environment variable access documented
技能调用	`NONE`	`READ`	✓ 一致	SKILL.md: mentions feishu_doc tool usage (create/append methods)
剪贴板	`NONE`	`NONE`	—	No clipboard access documented
浏览器	`NONE`	`NONE`	—	Mentions auto_open_chrome config but no actual browser automation code
数据库	`NONE`	`NONE`	—	No database access documented

目录结构

2 文件 · 7.5 KB · 368 行

Markdown 1f · 349L JSON 1f · 19L

├─ 📋 skill.json JSON 19L · 490 B

└─ 📝 SKILL.md Markdown 349L · 7.0 KB

安全亮点

✓ No malicious code or scripts present in the package

✓ No credential harvesting or environment variable access

✓ No network calls or data exfiltration mechanisms

✓ No obfuscated code or base64-encoded payloads

✓ No suspicious file access patterns (no ~/.ssh, ~/.aws, .env access)

✓ No supply chain risks (no external dependencies)

✓ Clean JSON metadata with proper MIT license