ClawHub
Back to skill

Security audit

A股数据结构化

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Eastmoney financial-data lookup tool that sends the user's query to the expected API and saves result files locally.

Install only if you are comfortable sending financial lookup queries to Eastmoney MX with your MX_APIKEY and retaining generated result files under MX_OUTPUT_DIR or the default ~/.codex/skills-output/mx_data/output path. Avoid putting confidential portfolio strategy or proprietary research intent into query text unless that is acceptable for your data-use policy.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)

Description-Behavior Mismatch

Medium
Confidence
81% confidence
Finding
The skill writes Excel and text artifacts to local disk as part of normal execution, which exceeds a narrow 'lookup and answer' expectation and creates persistence of queried financial data on the host. In an agent setting, undisclosed file creation can surprise users, leave sensitive query history behind, and enlarge the data exposure surface if the output directory is shared or monitored.

Description-Behavior Mismatch

Medium
Confidence
92% confidence
Finding
Saving the complete raw API response to disk is riskier than saving only the formatted user-facing result because it may contain extra metadata, identifiers, or fields the user did not request. This unnecessarily broadens local data retention and could expose provider response details or sensitive query-associated data to other local processes or future users of the system.

Missing User Warnings

Medium
Confidence
76% confidence
Finding
The code transmits the user's natural-language financial query and an API credential to a third-party service without any user-facing disclosure or consent mechanism in the skill itself. In an agent environment, undisclosed external transmission can expose sensitive prompts, portfolio identifiers, or proprietary research intents beyond what a user reasonably expects from a simple lookup action.

Missing User Warnings

Low
Confidence
74% confidence
Finding
The skill writes multiple files to disk without an explicit warning at the point of use, which can create unexpected local persistence of user queries and returned data. While not inherently malicious, silent file creation is a security and privacy concern in agent contexts because users may assume a lookup action is ephemeral.

External Transmission

Medium
Category
Data Exfiltration
Content
"toolQuery": tool_query
        }
        
        response = requests.post(self.BASE_URL, headers=headers, json=data, timeout=30)
        response.raise_for_status()
        return response.json()
Confidence
84% confidence
Finding
This skill sends user-supplied query content to an external network endpoint, which is an actual data egress behavior. In context, network access is expected for a real-time market data skill, so the issue is not the existence of the request itself but the privacy and trust implications of transmitting user input off-host to a third party.

Static analysis

No suspicious patterns detected.