ClawHub
Back to skill

Security audit

维加斯通道交易

Security checks for vulnerabilities and agentic risk

Overview

This finance trading skill is mostly purpose-aligned, but it asks for command execution and API-key-backed market access while its scope and execution boundaries are unclear.

Review this skill carefully before installing. It may be useful for market analysis, but only use it with read-only or least-privilege API keys, require explicit confirmation for any live-trading or file-writing step, and do not let it run commands unless you have defined exactly which commands are allowed.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Intent-Code Divergence

Medium
Confidence
84% confidence
Finding
The document claims strong originality/safety properties while elsewhere explicitly depending on third-party data APIs and external platform integrations. This kind of contradictory trust signaling can mislead reviewers or users into overestimating the skill’s isolation, provenance, or safety controls, which increases the chance that risky integrations are enabled without proper scrutiny.

Intent-Code Divergence

Medium
Confidence
91% confidence
Finding
The skill states that it is 'pure SKILL.md/no code execution needed' while also declaring the exec tool and describing command execution, file handling, networking, retries, and environment configuration. This mismatch is dangerous because users or automated systems may grant or invoke the skill under a lower-risk assumption, when it is actually capable of executing commands and interacting with external resources.

Vague Triggers

Medium
Confidence
78% confidence
Finding
The skill description is extremely broad and vague, covering generic automation, workflow, decision support, and multiple scenarios without clear activation boundaries. Overly broad scope increases the risk that an agent will apply the skill in unintended contexts, including contexts involving sensitive data, external actions, or financial decisions, without an explicit user intent check.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The skill exposes the exec capability in metadata, but the user-facing description does not clearly warn that it may execute commands or access the local/runtime environment. Hidden or under-disclosed execution capability is risky because users may provide data or approve the skill without understanding that it can perform system-level actions, especially in a finance-related workflow that also references API keys and external services.

Static analysis

No suspicious patterns detected.