中文 EN

扫描报告

扫描新文件

⚠ 低风险 — 风险评分 20/100
上次扫描：17 小时前 重新扫描
20 /100
alphapai-scraper
AlphaPai financial comment scraper and archival search tool
A legitimate, functional web scraping and archival skill for AlphaPai financial platform with minor documentation gaps but no malicious behavior detected.
技能名称alphapai-scraper
分析耗时77.6s
引擎pi
✓
可以安装
The skill is safe to use. Consider pinning dependency versions in setup.sh and documenting subprocess usage for osascript and openclaw agent calls in SKILL.md for full transparency.

安全发现 3 项

严重性 安全发现 位置
低危
Undeclared subprocess execution 文档欺骗
SKILL.md does not mention that run.py executes 'osascript' for macOS notifications, analyze.py invokes 'openclaw agent' subprocess for AI analysis, or publish_skill.py calls 'clawhub' CLI. These are legitimate tools but should be declared.
subprocess.run(['osascript', '-e', script], check=False, ...)
→ Add shell:WRITE to SKILL.md allowed-tools section, documenting osascript and openclaw agent usage.
 scripts/run.py:43
低危
Unpinned playwright dependency 供应链
setup.sh installs playwright without version pinning: 'pip3 install playwright --quiet'. This allows a malicious version to be pulled at install time.
pip3 install playwright --quiet
→ Pin to a known-safe version: pip3 install playwright==1.44.0 --quiet
 setup.sh:15
低危
Chrome profile directory access undeclared 敏感访问
The scraper accesses '~/Library/Application Support/Google/Chrome' for browser profile authentication. This is a sensitive path that should be declared in SKILL.md.
profile_user_data_dir: ~/Library/Application Support/Google/Chrome
→ Document browser:WRITE and the Chrome profile access in SKILL.md, as it implies access to all Chrome-stored credentials and sessions.
 config/settings.example.json:19
资源类型声明权限推断权限状态证据
文件系统 WRITE WRITE ✓ 一致 SKILL.md declares data storage; archive_store.py writes SQLite, raw, normalized …
网络访问 READ WRITE ✓ 一致 SKILL.md declares AlphaPai scraping; send_feishu.py POSTs to Feishu webhook (dec…
命令执行 NONE WRITE ✓ 一致 run.py:43 osascript subprocess; analyze.py:112 openclaw agent subprocess; publis…
环境变量 READ READ ✓ 一致 common.py:95-106 reads ALPHAPAI_* env vars; load_auth_bundle reads USER_AUTH_TOK…
技能调用 NONE WRITE ✓ 一致 analyze.py:112 invokes 'openclaw agent --message' as subprocess
剪贴板 NONE NONE —
浏览器 READ WRITE ✓ 一致 SKILL.md declares browser-based scraping; scraper.py uses Playwright with WRITE …
数据库 WRITE WRITE ✓ 一致 archive_store.py creates SQLite (FTS5) and Chroma vector index
3 项发现
🔗
中危 外部 URL 外部 URL
https://alphapai-web.rabyte.cn
config/settings.example.json:3
🔗
中危 外部 URL 外部 URL
https://alphapai-web.rabyte.cn/login
config/settings.example.json:4
🔗
中危 外部 URL 外部 URL
https://alphapai-web.rabyte.cn/reading/home/comment
config/settings.example.json:5

目录结构

19 文件 · 109.3 KB · 3476 行
Python 11f · 2929L Markdown 2f · 385L Shell 1f · 91L JSON 4f · 64L YAML 1f · 7L
├─ ▾ 📁 agents
│ └─ 📋 openai.yaml YAML 7L · 344 B
├─ ▾ 📁 config
│ ├─ 📋 cookies.example.json JSON 12L · 211 B
│ ├─ 🔑 credentials.example.json ⚠ JSON 4L · 70 B
│ ├─ 📋 settings.example.json JSON 43L · 1.2 KB
│ └─ 🔑 token.example.json ⚠ JSON 5L · 120 B
├─ ▾ 📁 scripts
│ ├─ 🐍 analyze.py Python 287L · 9.3 KB
│ ├─ 🐍 archive_store.py Python 748L · 24.5 KB
│ ├─ 🐍 bootstrap_session.py Python 84L · 3.0 KB
│ ├─ 🐍 common.py Python 320L · 9.8 KB
│ ├─ 🐍 init_config.py Python 94L · 2.6 KB
│ ├─ 🐍 package_skill.py Python 57L · 1.3 KB
│ ├─ 🐍 publish_skill.py Python 105L · 2.6 KB
│ ├─ 🐍 query_comments.py Python 228L · 7.9 KB
│ ├─ 🐍 run.py Python 177L · 5.8 KB
│ ├─ 🐍 scraper.py Python 755L · 24.6 KB
│ └─ 🐍 send_feishu.py Python 74L · 2.1 KB
├─ 📝 README.md Markdown 236L · 5.7 KB
├─ 🔧 setup.sh Shell 91L · 3.2 KB
└─ 📝 SKILL.md Markdown 149L · 4.8 KB

依赖分析 4 项

包名版本来源已知漏洞备注
playwright * pip 否 Version not pinned in setup.sh
chromadb * pip 否 Version not pinned; used for local vector search
sentence-transformers * pip 否 Version not pinned; loads bge-small-zh-v1.5 model locally
openclaw CLI system 否 Called via subprocess for AI summarization

安全亮点

✓ No evidence of credential exfiltration — tokens and cookies stay local to configured storage paths
✓ No obfuscation, base64 payloads, or anti-analysis techniques detected
✓ No network requests to IP addresses or suspicious domains beyond the declared AlphaPai platform and Feishu webhook
✓ package_skill.py correctly excludes all credential files before publishing (token, cookies, credentials, settings local files)
✓ Credential handling follows a clear, documented priority chain with no surprise collection
✓ No reverse shell, C2 communication, or data theft patterns found
✓ Code is clean, readable, and well-structured with no hidden functionality
✓ Normalize_cookies() in common.py filters cookie keys to a safe allowlist (name, value, domain, etc.)