Apex Crypto Intelligence

Security checks across malware telemetry and agentic risk

Overview

This crypto-analysis skill is mostly purpose-aligned, but it handles exchange API secrets and its verification mode falsely presents live third-party network calls as if nothing is transmitted.

Review before installing. Use only read-only exchange keys with trading and withdrawal permissions disabled, and assume running client.py may contact CoinGecko and exchange APIs even in its verification mode. Do not treat the verification script as an offline audit until that wording and behavior are fixed.

SkillSpector

By NVIDIA

Vulnerability Patterns

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
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (1)

Intent-Code Divergence

High

Confidence: 98% confidence
Finding: The verification path explicitly claims that 'nothing is transmitted' and 'no data was sent to any server', but it actually performs live requests to CoinGecko and multiple exchanges before printing the payload. This is dangerous because users may run verification mode expecting a safe offline inspection, while the script still discloses their IP, behavioral metadata, and potentially authenticated API-key usage to third-party services.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal