Back to skill

Security audit

Crypto Market Data

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed, read-only CoinGecko market-data client, though its manifest should be tightened to avoid stock or dividend questions being routed to a crypto-only tool.

Install this if you want crypto market data through AIsa/CoinGecko and are comfortable providing an AISA_API_KEY. Do not rely on it for equities, dividends, or traditional finance workflows; the publisher should narrow the manifest routing text to crypto-only use.

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 (2)

Description-Behavior Mismatch

Medium
Confidence
93% confidence
Finding
The manifest advertises the skill for stock analysis and dividend research even though the body later states it should not be used for equities or traditional finance. This can cause an agent to invoke the wrong skill in financial workflows, producing incorrect or inapplicable outputs that may mislead users making money-related decisions.

Vague Triggers

Medium
Confidence
95% confidence
Finding
Overly broad trigger text increases the chance that automated skill selection routes traditional-finance or dividend questions into a crypto-only data source. In agent ecosystems, bad routing is a security/reliability risk because it can generate authoritative-sounding but irrelevant financial answers, especially when users may trust automated research outputs.

Static analysis

No suspicious patterns detected.