Back to skill

Security audit

a-share-analysis

Security checks across malware telemetry and agentic risk

Overview

This is a coherent A-share analysis skill, but it should be reviewed because it gives actionable trading guidance while storing local history and using external services without tight user controls.

Install only if you are comfortable with an agentic stock tool that can produce trading recommendations, contact external financial/news services, and retain analysis history in your OpenClaw workspace. Treat outputs as informational, review generated files periodically, avoid using it for personalized investment decisions, and be cautious with the Firecrawl authentication helper and documented rm -rf cleanup commands. VirusTotal was pending, so this verdict rests on artifact behavior rather than malware telemetry.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
Findings (17)

subprocess module call

Medium
Category
Dangerous Code Execution
Content
}
        
        try:
            result = subprocess.run(
                [self.cli_path, "--status"],
                capture_output=True,
                text=True,
Confidence
89% confidence
Finding
result = subprocess.run( [self.cli_path, "--status"], capture_output=True, text=True, timeout=10 )

subprocess module call

Medium
Category
Dangerous Code Execution
Content
return False
            
            # 启动认证
            result = subprocess.run(
                [self.cli_path, "login", "--browser"],
                timeout=120
            )
Confidence
91% confidence
Finding
result = subprocess.run( [self.cli_path, "login", "--browser"], timeout=120 )

subprocess module call

Medium
Category
Dangerous Code Execution
Content
return None
        
        try:
            result = subprocess.run(
                [self.cli_path, "credit-usage", "--json"],
                capture_output=True,
                text=True,
Confidence
88% confidence
Finding
result = subprocess.run( [self.cli_path, "credit-usage", "--json"], capture_output=True, text=True, timeout=10 )

Intent-Code Divergence

Medium
Confidence
94% confidence
Finding
The code presents concrete investment recommendations, target/stop-loss prices, and position-sizing guidance, while simultaneously claiming the report is only for reference and not investment advice. This mismatch can mislead users into over-trusting the output and may create legal/compliance exposure, especially in a financial-analysis skill where users are likely to act on the recommendations.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The report states that the tool will automatically persist analysis results to multiple local files and update SESSION-STATE.md, but it does not clearly warn users that execution changes local state. This can cause unintended data retention, overwrite existing context, or leak sensitive analysis history into shared workspaces, especially in agent environments where users may assume a read-only analysis action.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The usage guide shows commands for running the analysis without warning that news sentiment features may contact external services such as Firecrawl and other financial data sources. In an agent skill context, undisclosed outbound requests can transmit stock queries, prompts, or related metadata to third parties, creating privacy, compliance, and unexpected network exposure risks.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The document includes destructive `rm -rf` cleanup commands without any explicit warning that the operation is irreversible or guidance to verify the target path before execution. In agent or automation contexts, users may copy these commands directly, creating a realistic risk of accidental data loss.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The trigger text is broad enough to activate on many generic stock- or market-related requests, increasing the chance the skill runs outside a narrowly intended context. In a financial-analysis skill, overbroad activation is risky because it can cause the agent to provide regulated or policy-sensitive investment guidance when the user did not explicitly request a bounded analytical workflow.

Natural-Language Policy Violations

High
Confidence
99% confidence
Finding
The skill description explicitly includes investment advice as a supported use case, which can steer the agent into producing actionable financial guidance. That is dangerous because users may rely on such outputs for real trading decisions, creating policy, compliance, and user-harm risk even if the surrounding text says the report is for reference only.

Natural-Language Policy Violations

High
Confidence
99% confidence
Finding
This section instructs the agent to produce concrete buy/hold/sell recommendations, crossing from analysis into prescriptive financial advice. In context, that makes the skill more dangerous because it operationalizes the advisory behavior as part of the standard workflow rather than an incidental example.

Natural-Language Policy Violations

High
Confidence
98% confidence
Finding
The example scenario asking whether a stock is worth buying normalizes actionable investment-decision support and primes the agent to answer in a recommendation style. Because examples heavily influence runtime behavior, this context increases the likelihood of policy-violating outputs and real-world financial harm.

Vague Triggers

Medium
Confidence
94% confidence
Finding
The trigger text is broad enough to activate on general investing or market-analysis requests, which can cause the skill to run in situations the user did not specifically intend. In this skill, that matters because invocation can lead to external news scraping, generation of investment recommendations, and persistence of analysis data, expanding both privacy and autonomy risk.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill describes storing analysis history in persistent local files, including per-stock histories and session state, but does not clearly disclose to users that their prompts, derived recommendations, and related context may be retained. This creates a privacy and data-retention risk because sensitive investment interests and decision history could be stored longer than users expect and later exposed to other tools, users, or local file access.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The Firecrawl-based news feature implies outbound network requests using user-selected stock identifiers and names, but the skill does not explicitly warn that this context may be sent to a third-party service. That omission is dangerous because users may not realize their financial interests or research targets are being disclosed externally, and third-party processing introduces additional confidentiality and compliance risk.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The script uses plain HTTP for external API endpoints and even sets an HTTP Referer, so stock symbols and returned market data travel without transport-layer protection. This enables man-in-the-middle interception or tampering on untrusted networks, which can cause users to receive falsified technical analysis inputs and make bad trading decisions.

Tool Parameter Abuse

High
Category
Tool Misuse
Content
### 清理报告
```bash
# 删除某支股票的所有报告
rm -rf a-share-reports/600519/

# 删除所有报告
rm -rf a-share-reports/*
Confidence
96% confidence
Finding
rm -rf a-share-reports/600519/

Tool Parameter Abuse

High
Category
Tool Misuse
Content
rm -rf a-share-reports/600519/

# 删除所有报告
rm -rf a-share-reports/*
```

---
Confidence
99% confidence
Finding
rm -rf a-share-reports/

VirusTotal

61/61 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.