Security audit

加密组合跟踪

Security checks for vulnerabilities and agentic risk

Overview

This skill mostly describes crypto portfolio analysis, but it also introduces automated trade execution and broad command/file capabilities without clear limits or user confirmation.

Review this carefully before installing. It may be useful for crypto portfolio analysis, but do not give it exchange trading keys, wallet-signing access, or command authority unless you intend to allow financial actions and can enforce explicit confirmation for every trade or file/system change.

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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

Findings (6)

Intent-Code Divergence

High

Confidence: 95% confidence
Finding: The skill is presented as a portfolio-tracking tool, but the documentation later introduces automatic trading execution. This scope expansion is dangerous because it can cause the agent or user to authorize financial actions they did not expect, especially in a high-risk crypto context where trades can immediately move funds.

Description-Behavior Mismatch

High

Confidence: 96% confidence
Finding: The manifest describes a tracking/analysis skill, but the body documents automated trade execution, creating a mismatch between declared and actual behavior. Such deceptive or sloppy scoping undermines user consent and can bypass safety expectations for a supposedly read-oriented financial tool.

Context-Inappropriate Capability

High

Confidence: 94% confidence
Finding: Automatic trade execution is not justified by the stated purpose of portfolio tracking and materially increases the skill's risk profile. In a crypto setting, unnecessary execution capability can lead to unauthorized trades, financial loss, or abuse if invoked with malformed or adversarial inputs.

Missing User Warnings

High

Confidence: 95% confidence
Finding: The markdown advertises automated trading execution without a clear, prominent warning that the skill may trigger real financial actions. In a cryptocurrency context, lack of explicit warning and consent is especially dangerous because mistakes or abuse can cause immediate irreversible losses.

Missing User Warnings

Medium

Confidence: 84% confidence
Finding: The skill claims file write capability in documentation without a clear warning about when data may be modified or where outputs may be written. Undisclosed write behavior can surprise users, overwrite files, or be abused to persist untrusted content in the workspace.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The skill documents command execution capability without a clear warning or constrained execution model. Because the skill also has the exec tool declared, ambiguous command use can expose the environment to command injection, unsafe shell actions, or misuse beyond simple portfolio analysis.

Static analysis

No suspicious patterns detected.