ClawHub

CoinGecko API

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed CoinGecko API reference skill for live crypto data, with no executable code or hidden local-system behavior found.

Install this if you want your agent to use CoinGecko/GeckoTerminal for live crypto data. Provide only a CoinGecko API key you are comfortable using with the agent, expect requests to CoinGecko domains, and treat on-chain token links and wallet-level analytics as unverified or privacy-sensitive when sharing results.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

Low
Confidence
88% confidence
Finding
The guidance explicitly tells the agent to fall back to unauthenticated keyless access if no valid key is available, but it does not require user consent or warning about the privacy, reliability, and control tradeoffs. In an agent setting, that can cause unintended external requests over a shared/public access tier and may surprise users who expected authenticated, account-scoped behavior.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
This documentation exposes endpoints that return wallet addresses, labels, explorer links, holder concentration, and trader PnL/profiling data without any caution that these fields are privacy-sensitive and can be used for deanonymization, profiling, or targeted surveillance of blockchain users. In an agent skill context, the absence of guardrails increases the chance the agent will surface or normalize wallet-level intelligence to end users without minimization or consent considerations.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The file correctly notes that token metadata may be sourced on-chain and not vetted, but it omits the more actionable warning that embedded websites and social links may be malicious, deceptive, or unsafe to visit. In an agent environment, this can lead the system to present or recommend untrusted URLs as if they were authoritative, creating phishing and social-engineering risk.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal