Security audit

数据质量检查基础版

Security checks for vulnerabilities and agentic risk

Overview

This looks like a data-quality skill, but it under-explains network/API use and includes broad environment and file/command authority.

Review this before installing if your datasets, environment variables, or API keys are sensitive. Use it only on explicitly selected local files, avoid running the broad environment-variable check, and do not provide callback URLs or external API credentials unless you understand what data may leave your machine.

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 (5)

Description-Behavior Mismatch

Medium

Confidence: 95% confidence
Finding: The skill claims that FREE version data is stored locally and implies privacy safety, yet elsewhere states that some functionality requires network access and an LLM API. This can mislead users into supplying sensitive financial or proprietary datasets under false assumptions about data residency and external transmission.

Context-Inappropriate Capability

Medium

Confidence: 97% confidence
Finding: The skill instructs users or the agent to enumerate environment variables matching API/KEY/TOKEN/SECRET, which exposes the presence and names of sensitive credentials unrelated to basic data-quality checking. In an agent context with exec capability, this broad secret discovery behavior can enable credential harvesting or accidental leakage into logs and outputs.

Context-Inappropriate Capability

Medium

Confidence: 90% confidence
Finding: The manifest grants generic exec, write, read, and glob access even though the described function is a lightweight personal data-quality checker. Excessive capability scope increases the blast radius of prompt misuse, enabling arbitrary command execution or file access beyond what is needed for simple validation tasks.

Vague Triggers

Medium

Confidence: 85% confidence
Finding: The trigger conditions are broad enough to activate the skill for general data analysis, reporting, or visualization tasks beyond strict data-quality checks. In an agent environment, overbroad invocation can cause unnecessary file access, command execution, or unintended handling of sensitive datasets when a narrower tool should have been used.

Vague Triggers

Medium

Confidence: 83% confidence
Finding: The natural-language invocation guidance is so generic that the skill may be selected for vague requests without adequate scoping. Because the skill has exec and file capabilities, accidental invocation raises the risk of unnecessary system interaction and exposure of local data during ambiguous workflows.

Static analysis

No suspicious patterns detected.