Security audit

Graph Query Tool Free

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a blockchain data-query helper, but it asks for broader agent authority and gives underspecified network and write instructions that users should review before installing.

Install only if you are comfortable with an agent making external blockchain-data requests and potentially sending addresses, protocol names, and query details to third-party services. Avoid using sensitive or personally linkable wallet addresses unless needed, and require explicit confirmation before any save, export, import, modify, package-install, or shell/network diagnostic action.

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 (5)

Description-Behavior Mismatch

Medium

Confidence: 89% confidence
Finding: The skill is presented as a read-oriented blockchain query tool, but the documentation explicitly advertises create/modify/import/export/save operations via generic parameters. In an agent environment with write and exec capabilities enabled, this ambiguity can broaden the agent's behavior beyond passive querying and cause unintended file writes, imports, or state-changing actions that the user did not expect.

Intent-Code Divergence

Medium

Confidence: 82% confidence
Finding: The documentation claims the free version never requests a wallet, yet the quota example requires a sender parameter that looks like a wallet address. This inconsistency can mislead users and agents into disclosing blockchain identifiers or reusing sensitive addresses under the assumption that no wallet-related data is needed.

Intent-Code Divergence

Low

Confidence: 77% confidence
Finding: The error handling instructs the agent to run network diagnostic commands such as ping even though the skill is described as a blockchain query tool. In environments where exec is available, this can trigger unnecessary command execution, expand system interaction beyond the skill's purpose, and expose host/network metadata.

Vague Triggers

Medium

Confidence: 80% confidence
Finding: The trigger condition says to use the skill for broad tasks like data analysis, report generation, statistical insight, and visualization, which extends far beyond blockchain query routing. Overbroad activation criteria can cause an agent to invoke this networked skill in unrelated contexts, increasing unintended data disclosure to third-party services and misuse of exec/write-enabled capabilities.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill does not prominently warn that use of the skill may send user prompts, addresses, protocol names, and query parameters to external third-party data services. In a security-sensitive agent setting, lack of disclosure increases the risk of silent data exfiltration and prevents informed consent before network access occurs.

Static analysis

No suspicious patterns detected.