ClawHub
Back to skill

Security audit

Tushare

Security checks across malware telemetry and agentic risk

Overview

This Tushare skill mostly matches its finance-data purpose, but it also exposes account portfolio save/delete APIs without clear confirmation rules.

Install only if you are comfortable giving the skill access to your Tushare token for market-data retrieval. Treat exports and cache files as user-directed actions, and do not allow portfolio save/delete operations unless you explicitly requested and confirmed them. Static scan was clean and VirusTotal supplied no negative telemetry, so the review concern is scope and confirmation rather than evidence of malware.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The interface list exposes state-changing portfolio management operations (p_save, p_delete, p_get/p_list) inside a skill described as research-oriented data retrieval and analysis. This expands capability from read-only market research into write/delete actions on user-managed objects, creating an authorization and intent-confusion risk where natural-language requests or prompt injection could trigger persistent changes the user did not clearly approve.

Vague Triggers

Medium
Confidence
94% confidence
Finding
The trigger guide uses broad, everyday finance phrases such as asking to 'quickly research' a stock or 'pull some data', which can cause the skill to activate when the user did not explicitly intend to invoke external data retrieval or file export behavior. Over-broad activation increases the chance of unnecessary network access, unintended use of credentials, and surprising actions in mixed-conversation contexts.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The quick-rule section says that if a user says very generic things like 'look at the trend', 'check financials', or 'compare companies', the skill should activate before considering more specific scope constraints. This can lead to accidental invocation and unnecessary external API use, especially when similar phrases appear in brainstorming, translation, or hypothetical discussion rather than an actual request to run the skill.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.