Security audit

代码

Security checks across malware telemetry and agentic risk

Overview

This skill appears to be a code-explanation helper, but it asks for command execution and file-writing authority while also describing broader code generation, review, testing, deployment, and batch-processing uses.

Review this skill carefully before installing. It may be reasonable if you want a broad coding assistant, but it is over-permissioned and under-scoped for a simple code-explanation skill; prefer a version that removes exec/write permissions or clearly limits them to user-approved, explanation-related actions.

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)

Description-Behavior Mismatch

High

Confidence: 94% confidence
Finding: The skill is presented as a code-explanation tool, but its metadata and prose broaden scope into code generation, review, testing, deployment, and refactoring. This scope drift can cause an agent to invoke the skill in far more privileged or safety-sensitive contexts than users expect, enabling unintended handling of code modification or operational workflows.

Context-Inappropriate Capability

High

Confidence: 97% confidence
Finding: Declaring exec and write for a skill whose stated purpose is explanation violates least privilege and expands the blast radius if the skill is invoked on untrusted input. Even absent explicit exploit logic, these capabilities allow file modification and command execution in contexts where only read access should be necessary.

Intent-Code Divergence

Medium

Confidence: 84% confidence
Finding: The input contract says the skill explains code, but the output schema is structured as a grading/audit report. This mismatch can mislead orchestration logic and downstream consumers into treating the skill as a reviewer or scorer, creating unsafe task routing and increasing the chance of the skill being used beyond its stated purpose.

Intent-Code Divergence

Medium

Confidence: 90% confidence
Finding: The FAQ explicitly states that the skill supports code generation, directly contradicting the advertised explanation-only purpose. This creates a clear path for overbroad invocation and may cause users or agents to rely on the skill for code-producing tasks that require stronger controls and more precise safety boundaries.

Vague Triggers

Medium

Confidence: 80% confidence
Finding: The invocation guidance is broad enough to match generic coding requests such as generation, debugging, testing, and deployment. Overbroad activation criteria increase the chance that an agent will select this skill in situations outside its safe or intended scope, especially given its declared exec/write permissions.

Vague Triggers

Medium

Confidence: 79% confidence
Finding: The description lacks precise activation constraints and uses broad, promotional language that does not clearly bound behavior. Ambiguous trigger scope is risky because agent selection systems may infer capabilities beyond what is safe, especially when combined with powerful tools and contradictory documentation.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.