Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

Eastmoney Trading

v3.2.0

东方财富证券交易技能,支持自动登录、持仓查询、持仓分析、条件选股、买入、卖出、撤单、委托查询、资金查询等完整交易功能。使用 CDP 连接浏览器,支持验证码自动识别。⚠️ 高风险操作,需谨慎使用。

1· 874·7 current·7 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Suspicious
high confidence
!
Purpose & Capability
The code and SKILL.md implement trading (login, buy/sell, queries) which matches the description. However the registry metadata lists no required credentials or config paths while the SKILL.md and script require EASTMONEY_ACCOUNT/EASTMONEY_PASSWORD (and optionally EASTMONEY_CDP_URL) and may read a local .env — this mismatch between declared requirements and actual needed secrets/config is incoherent.
!
Instruction Scope
Runtime instructions and the script instruct the agent/user to: configure OpenClaw openclaw.json, run/attach to a remote CDP URL, use environment variables or a local .env for account/password, install Tesseract, run playwright-based automation, save screenshots and logs. The script also sends captcha images to an external LLM OCR endpoint when an API key is provided. Reading .env, creating logs/screenshots, and calling third-party OCR services are outside the narrow surface of 'just issuing API calls to the brokerage' and are worth flagging.
Install Mechanism
There is no platform install spec (instruction-only + packaged Python script). A requirements.txt is provided (playwright, Pillow, pytesseract, requests, pandas). Playwright may require browser binaries or post-install steps on the host. There are no suspicious remote download URLs in the install, but installing Playwright can pull additional binaries — verify before running.
!
Credentials
The skill requires sensitive credentials (EASTMONEY_ACCOUNT/EASTMONEY_PASSWORD) and optionally reads API keys (BAILIAN_API_KEY, DASHSCOPE_API_KEY, ALIBABA_CLOUD_API_KEY) for LLM OCR — yet the registry metadata declared no required env vars/primary credential. That mismatch is problematic. Also the skill writes logs/screenshots containing financial details (and possibly captcha images) to disk by default; those artifacts may expose credentials or account activity if not protected.
Persistence & Privilege
always:false and user-invocable:true (normal). The skill writes logs/screenshots under its own directory (expected). Autonomous model invocation is allowed by default; because this skill can perform real trades, consider the extra risk if you permit autonomous invocation — combine that with the other concerns before allowing autonomous runs.
What to consider before installing
Key things to check before installing or running: - The registry metadata does NOT list the env vars the code needs. Expect to set EASTMONEY_ACCOUNT and EASTMONEY_PASSWORD (and possibly EASTMONEY_CDP_URL). Treat those as highly sensitive; use a test account first and never upload real credentials to cloud services. - The script can send captcha screenshots to a third-party LLM OCR endpoint (dashscope.aliyuncs.com) when an API key (BAILIAN_API_KEY, DASHSCOPE_API_KEY, or ALIBABA_CLOUD_API_KEY) is present. If you don't want images/leaked data sent externally, do not set those API keys and audit the code path that calls recognize_captcha_with_llm. - The skill saves logs and screenshots locally by default. Those files can contain sensitive financial data and images of UI screens; secure or regularly purge the logs/screenshots directory, or modify the script to avoid saving sensitive artifacts. - Playwright and pytesseract are required; installing Playwright can pull browser binaries. If you rely on the claimed 'CDP only' mode, verify the code actually avoids downloading/launching new browsers and only connects over CDP. - Run the script in an isolated environment (VM/container) with a test account first to observe network egress (which domains are contacted) and file writes. Review/grep the full script for any additional outgoing endpoints before giving it real account credentials. - If you want to use LLM OCR, prefer using your own trusted API keys and confirm the provider's privacy policy; otherwise disable LLM OCR and rely on local Tesseract. Overall: the functionality aligns with the stated purpose, but undeclared env vars and third-party OCR/network behavior plus default logging of sensitive data are important red flags — proceed only after verifying and hardening those points.

Like a lobster shell, security has layers — review code before you run it.

latestvk97f2z4s75pvwkcej5bj2na8v982waxk

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments