ETH价格预测

• Outcome: Review
• Latest audit: May 28, 2026 · 1:47 AM UTC
• Version: 1.0.0

Overview

This paid ETH prediction skill has real billing behavior, but its payment identity, hardcoded billing key, and advertised features do not line up safely with what users are told.

Review carefully before installing or running. Do not deposit funds or rely on the predictions unless the author removes the embedded billing key, uses a stable authenticated billing identity, adds explicit charge confirmation, and either implements or removes the claimed on-chain and ETF data features.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (5)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 85% confidence
• Finding: The skill declares access to an environment variable and appears to require networked behavior, but does not explicitly declare permissions for those capabilities. Hidden or undeclared capability use reduces transparency for users and reviewers, especially in a paid skill that interacts with external services, and can conceal data access or outbound requests beyond what the documentation suggests.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 97% confidence
• Finding: The skill's documented purpose materially differs from its actual behavior: it charges via an external billing API, may generate top-up links, and claims analytics sources/features that are not actually implemented. This is dangerous because users may invoke a paid financial tool under false assumptions about what data is used and what actions occur, leading to deceptive charges, trust abuse, and risky financial decision-making.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: A paid charge is mentioned in the document, but the finding indicates the skill description does not present a clear up-front warning that every invocation incurs a cost. In a transactional skill, insufficient charge disclosure can cause users to trigger billable actions unintentionally, especially when commands are presented as normal utility operations.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The code sends user_id, skill_id, and billing amount to an external billing endpoint with no visible consent check, disclosure, or minimization. Even if billing is expected, silently transmitting identifiers and transaction metadata to a third party can expose user privacy and create compliance issues, especially if this occurs automatically on invocation.

Missing User Warnings

Medium

• Confidence: 99% confidence
• Finding: A live-looking API key is hardcoded as a fallback secret in source code. If this file is exposed, shared, or bundled, attackers can extract the credential and perform unauthorized billing operations, generate payment links, or abuse the provider account.

VirusTotal

66/66 vendors flagged this skill as clean.