Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Human Stock Helper
v1.0.0为股票交易者提供理性分析、技术计算和策略执行提醒。基于AKShare、yfinance、pandas-ta等工具,帮助用户冰冷、果决地执行交易决策。
⭐ 2· 257·1 current·1 all-time
byMaxHou_12@maxhou-infinity
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The name/description match the included Python tools for position/strategy management and analysis. However, the SKILL.md and code disagree about installation requirements (SKILL.md lists Python packages), and the code hard-codes a developer-specific filesystem path (/Users/maxhou/Desktop/openclawmax/memory) which is unexpected for a generic skill and not declared in the registry metadata.
Instruction Scope
SKILL.md instructs the agent to call external tools/services (Tavily via 'mcporter', AKShare, backtester scripts) and to read/write various 'memory/..' JSON files. The code indeed reads/writes JSON files, but uses an absolute developer path instead of the documented 'memory/' relative paths. SKILL.md references scripts/backtest_strategy.py and mcporter calls that are not present or not documented in the registry (missing script and missing declaration of credentials), meaning the agent could attempt to call external services or paths not under the user's control.
Install Mechanism
Registry metadata indicated no install spec but SKILL.md contains an install block recommending pip installs of akshare, yfinance, pandas, pandas-ta (the pip install list in SKILL.md omits pandas while py_packages lists it) — inconsistent. Installing these pip packages is standard but will pull network-accessible code; there is no unusual or remote-URL download beyond normal pip packages.
Credentials
The skill declares no required env vars or config paths, yet the code hard-codes a full absolute path on the developer's machine for all storage. The skill also directs data to external services (Tavily via mcporter, yfinance/AKShare network calls) without declaring credentials or endpoints. This mismatch means the skill may read/write/wipe files in an unexpected location and may transmit user-provided queries/holdings to external services not enumerated in the registry metadata.
Persistence & Privilege
The skill is not marked always:true and does not request special platform-wide privileges. It writes and reads files under a specific path, but it does not modify other skills or system-wide agent settings. The main concern is the hard-coded filesystem path and undeclared external calls, not elevated platform privileges.
What to consider before installing
This skill is plausibly what it says (a trading helper), but there are several red flags you should address before installing or running it:
- Hard-coded path: The code reads/writes memory files at /Users/maxhou/Desktop/openclawmax/memory. That is a developer-specific absolute path and could cause the skill to fail, overwrite, or attempt to read data in that location on your machine. Ask the author to make the storage directory configurable (or relative) and to not hard-code a home-folder path.
- Metadata mismatch: SKILL.md recommends pip packages (akshare, yfinance, pandas, pandas-ta) but the registry metadata claims no install spec. The pip list in SKILL.md and the metadata are inconsistent (pandas appears in py_packages but not in the pip packages list). Installing packages will allow network access; review them and install in a virtualenv or sandbox.
- Undeclared external calls: SKILL.md instructs use of 'mcporter call tavily.*' (an external service) and a backtest script path that is not included. The skill may send queries (which could include your holdings or trade notes) to Tavily or other external services. Confirm what data is sent and whether credentials are required and declared.
- Missing/incorrect files: SKILL.md references scripts/backtest_strategy.py and some SKILL.md metadata differs from the included files. Confirm the repository includes all runtime scripts and that their behavior matches the documentation.
- Run safely: If you want to try it, run inside an isolated environment (container or VM), and create a dedicated empty directory to map as the memory directory; inspect/modify the code to use a safe path before executing. Backup any important files first.
- Ask the maintainer to: (1) remove hard-coded absolute paths or expose a configuration/env var for MEMORY_DIR, (2) declare install steps and required credentials clearly in registry metadata, (3) include or remove references to missing scripts/tools (backtest script, mcporter/Tavily), and (4) document what external endpoints will receive user data. If the author cannot justify these mismatches, treat the skill as unsafe to run with real account or sensitive data.Like a lobster shell, security has layers — review code before you run it.
latestvk97cynev38cm2f4c9wp75t0tbd82ksmp
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
📈 Clawdis
Binspython3
