Install untrusted source
Warn
- Finding
- Install source points to URL shortener or raw IP.
A Share Guard
Security checks across static analysis, malware telemetry, and agentic risk
Overview
This skill mainly analyzes public stock data and shows no credential theft or destructive behavior, but users should review its script execution, hard-coded browserless IP, and incomplete monitoring/data features before relying on it.
Before installing, confirm you are comfortable allowing the skill to run its bundled Python scripts and access public finance websites. Replace the 192.168.3.120 browserless URL with your own trusted endpoint, verify the data sources, and treat the generated stock-risk report as non-advisory information rather than a trading instruction.
Static analysis
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
What this means
The agent may run local scripts and fetch web pages when the skill is invoked.
Why it was flagged
The skill asks for web access and local command execution. That is aligned with public data collection and running bundled Python analysis scripts, but exec remains broad authority.
Skill content
requires:\n tools: ["web_search", "web_fetch", "exec"]
Recommendation
Use it in an environment where local script execution is acceptable, and review or approve commands if your OpenClaw setup supports approval prompts.
What this means
If copied as-is, browser automation could connect to an unintended LAN service or receive untrusted page data.
Why it was flagged
The template points browserless/CDP integration at a raw private IP; scripts/f10_scraper.py also hard-codes the same URL. Browserless is purpose-aligned, but this endpoint has unclear provenance for other users.
Skill content
"cdpUrl": "http://192.168.3.120:3000",\n "attachOnly": true
Recommendation
Replace the CDP URL with your own trusted browserless endpoint, preferably authenticated, or remove the browserless config until needed.
What this means
If implemented by the platform or future code, the skill may continue checking stocks on a schedule after the initial command.
Why it was flagged
The documentation advertises recurring daily scans/watchlist behavior, but the artifacts do not show the scheduler, storage location, retention, or removal controls.
Skill content
| `/guard-watch [代码]` | 添加监控 | 每日自动扫描 |
Recommendation
Use watch mode only intentionally and confirm how to list, remove, or disable monitored stocks.
What this means
Reports may look authoritative even when some financial indicators are incomplete or unavailable.
Why it was flagged
Some core risk inputs are still placeholders/TODOs, while the SKILL.md markets expert-level financial diagnostics and gives action-style outputs. This is a reliability/trust caveat rather than evidence of malicious behavior.
Skill content
# 简化版：返回模拟数据（实际需要 browserless 抓取 F10 页面）\n # TODO: 实现 browserless CDP 抓取
Recommendation
Treat results as informational only, verify figures from primary financial sources, and do not rely on the skill alone for trading decisions.