中文 EN

扫描报告

扫描新文件

✓ 低风险 — 风险评分 5/100
上次扫描：1 天前 重新扫描
5 /100
memory-lancedb-pro
Production-grade long-term memory system (v1.1.0-beta.8) for OpenClaw AI agents. Provides persistent, intelligent memory storage using LanceDB with hybrid vector + BM25 retrieval, LLM-powered Smart Extraction, Weibull decay lifecycle, and multi-scope isolation.
Pure documentation skill (Markdown only) providing installation and configuration guidance for a memory plugin. No executable code, scripts, or binaries present. Minor concern: remote script download pattern in documentation.
技能名称memory-lancedb-pro
分析耗时40.5s
引擎pi
✓
可以安装
Safe to use. No executable components. The documented `curl|bash` remote script pattern is standard practice and clearly disclosed. No action required.

安全发现 2 项

严重性 安全发现 位置
低危
Remote script download pattern in documentation 供应链
The Quick Install section references downloading and executing a shell script from GitHub raw content: `curl -fsSL https://raw.githubusercontent.com/CortexReach/toolbox/main/memory-lancedb-pro-setup/setup-memory.sh -o setup-memory.sh && bash setup-memory.sh`. This is a documented pattern with `--dry-run` and `--selfcheck-only` options, but the remote script is not reviewed in this package.
curl -fsSL https://raw.githubusercontent.com/CortexReach/toolbox/main/memory-lancedb-pro-setup/setup-memory.sh -o setup-memory.sh
→ Review the external script at the GitHub URL before executing it. Use the `--dry-run` flag to preview actions first.
 SKILL.md:686
低危
Misleading Iron Rule references dangerous command 文档欺骗
The 'Iron Rules for AI Agents' section includes `rm -rf /tmp/jiti/` as a required step. While the target `/tmp/jiti/` is a legitimate cache directory (not root), the prefix `rm -rf` with a glob-like pattern is an IOC that was flagged by pre-scan. This is benign in context but creates false-positive security alerts.
rm -rf /tmp/jiti/
→ Use `rm -rf ~/.cache/jiti/` or a more specific path to reduce confusion and avoid false-positive security flags.
 SKILL.md:697
资源类型声明权限推断权限状态证据
文件系统 READ READ ✓ 一致 Markdown documentation files only; no file write operations
网络访问 READ READ ✓ 一致 Only URL references for external services (jina.ai, openai.com, GitHub) — all de…
命令执行 NONE NONE — No executable scripts present; shell commands appear only as documentation
环境变量 NONE NONE — No environment variable access
技能调用 READ READ ✓ 一致 Skill is itself a documentation resource
剪贴板 NONE NONE — No clipboard access
浏览器 NONE NONE — No browser automation
数据库 NONE NONE — No database access; documents LanceDB configuration but does not connect
1 严重 18 项发现
💀
严重 危险命令 危险 Shell 命令
rm -rf /
SKILL.md:697
🔗
中危 外部 URL 外部 URL
https://claude.ai/code
README.md:28
🔗
中危 外部 URL 外部 URL
https://openclaw.ai
README.md:32
🔗
中危 外部 URL 外部 URL
https://storage.ko-fi.com/cdn/kofi2.png?v=3
README.md:229
🔗
中危 外部 URL 外部 URL
https://ko-fi.com/aila
README.md:229
🔗
中危 外部 URL 外部 URL
https://jina.ai/api-key
SKILL.md:29
🔗
中危 外部 URL 外部 URL
https://platform.openai.com/api-keys
SKILL.md:29
🔗
中危 外部 URL 外部 URL
https://cloud.siliconflow.cn/account/ak
SKILL.md:38
🔗
中危 外部 URL 外部 URL
https://ollama.com/download
SKILL.md:62
🔗
中危 外部 URL 外部 URL
https://api.jina.ai/v1/embeddings
SKILL.md:88
🔗
中危 外部 URL 外部 URL
https://api.siliconflow.com/v1/rerank
SKILL.md:104
🔗
中危 外部 URL 外部 URL
https://api.jina.ai/v1
SKILL.md:162
🔗
中危 外部 URL 外部 URL
https://api.jina.ai/v1/rerank
SKILL.md:186
🔗
中危 外部 URL 外部 URL
http://192.168.1.100:11434/v1
SKILL.md:333
🔗
中危 外部 URL 外部 URL
https://dashscope.aliyuncs.com/compatible-mode/v1
SKILL.md:1223
🔗
中危 外部 URL 外部 URL
https://dashscope.aliyuncs.com/compatible-api/v1/reranks
SKILL.md:1236
🔗
中危 外部 URL 外部 URL
https://api.voyageai.com/v1/rerank
SKILL.md:1252
🔗
中危 外部 URL 外部 URL
https://api.pinecone.io/rerank
SKILL.md:1253

目录结构

3 文件 · 76.6 KB · 1951 行
Markdown 3f · 1951L
├─ ▾ 📁 references
│ └─ 📝 full-reference.md Markdown 310L · 9.7 KB
├─ 📝 README.md Markdown 233L · 9.8 KB
└─ 📝 SKILL.md Markdown 1408L · 57.0 KB

安全亮点

✓ Pure documentation skill — no executable code, scripts, or binaries
✓ All shell commands appear only as documentation examples, not as embedded execution
✓ API key verification uses read-only HTTP GET requests to service endpoints
✓ No credential harvesting — only user-provided keys via config files
✓ No obfuscation (no base64, eval, atob, or encoded strings)
✓ No sensitive file access (no ~/.ssh, ~/.aws, .env reads)
✓ No data exfiltration, reverse shell, or C2 communication patterns
✓ No hidden functionality — documentation matches stated behavior
✓ Remote script has `--dry-run` and `--selfcheck-only` safety options
✓ Full source file map and technical details transparently disclosed