Security audit

能力扩展工具免费版

Security checks across malware telemetry and agentic risk

Overview

This knowledge lookup skill may be useful, but it asks for broad file and shell authority and sends queries to an external service without clear enough limits.

Review this before installing. Use it only if you are comfortable with the agent sending lookup queries, and possibly surrounding context, to an external knowledge service. Avoid using it with secrets, proprietary code, private documents, or account data unless the provider’s endpoint, retention, and logging policies are clear. Grant file and shell tools only in a constrained workspace because those permissions are broader than the lookup purpose requires.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
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

Findings (5)

Description-Behavior Mismatch

Medium

Confidence: 93% confidence
Finding: The skill is presented as a narrowly scoped external knowledge retrieval tool, yet it also advertises broad file processing and shell execution capabilities. That mismatch expands the effective attack surface and can mislead users or orchestrators into granting permissions far beyond what is necessary, increasing the risk of unintended file access or command execution.

Context-Inappropriate Capability

Medium

Confidence: 88% confidence
Finding: Declaring code-related capability for a skill whose stated purpose is knowledge retrieval violates least privilege. In agent environments, unnecessary execution capability can enable prompt-triggered shell activity or code generation/execution pathways unrelated to the user’s intent.

Intent-Code Divergence

Medium

Confidence: 86% confidence
Finding: The documentation claims the skill is stateless and does not persist data, but elsewhere it references create/export operations and file handling. These contradictions can cause operators to trust the skill with sensitive data under false assumptions, creating privacy and integrity risks if content is actually written, exported, or retained.

Vague Triggers

Medium

Confidence: 81% confidence
Finding: The trigger guidance is broad enough that the skill may activate on many vague requests involving uncertainty or recency. In practice, this can cause unnecessary transmission of user prompts to an external service and may trigger higher-risk capabilities more often than users expect.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill describes authenticated external requests to a third-party knowledge service but does not clearly warn users that their inputs may leave the local environment. This creates a meaningful data exfiltration and privacy risk, especially because users may include proprietary code, credentials, or internal documents in troubleshooting or coding prompts.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.