Security audit

代码分析工具免费版

Security checks across malware telemetry and agentic risk

Overview

The skill is mostly a local analysis prompt, but it requests broad command and file capabilities and contains inconsistent local-only, API, network, and write-operation claims.

Review this skill before installing. It may be useful as a structured analysis template, but only enable it in an environment where command execution and possible file writes are acceptable, and avoid giving it sensitive code or documents unless you are satisfied that any network/API behavior is disabled or explicitly controlled.

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 (10)

Description-Behavior Mismatch

High

Confidence: 95% confidence
Finding: The skill is presented as a local structured analysis/reporting tool, yet the documentation later advertises external API integration and network communication. This mismatch can cause users and calling agents to grant trust or permissions under false assumptions, enabling unexpected data egress or capability expansion beyond the declared purpose.

Description-Behavior Mismatch

Medium

Confidence: 88% confidence
Finding: Documentation claims file write capabilities even though the skill is primarily described as an analysis tool that produces conclusions rather than modifying the local system. That discrepancy increases the risk of unauthorized or surprising file changes, especially when users invoke the skill for read-only analysis tasks.

Context-Inappropriate Capability

High

Confidence: 97% confidence
Finding: System command execution is a powerful capability that is not justified by the stated purpose of structured analysis. In a skill that accepts arbitrary input and natural-language instructions, unnecessary exec access materially raises the risk of command injection, destructive local actions, or unintended environment access.

Context-Inappropriate Capability

High

Confidence: 93% confidence
Finding: The skill claims local/offline structured analysis, but also documents API/network access without showing why outbound communication is needed. This creates a path for sensitive user-provided data to leave the local environment unexpectedly, which is especially dangerous for code, reports, or internal documents submitted for analysis.

Intent-Code Divergence

Critical

Confidence: 99% confidence
Finding: The documentation explicitly states that the tool works locally and does not require networking, but elsewhere describes network communication and API calling. This contradiction is highly dangerous because users may submit sensitive local data under the assumption that it never leaves the machine, when the skill text leaves room for outbound transmission.

Intent-Code Divergence

Medium

Confidence: 90% confidence
Finding: The skill states that free-version results are only manually saved or copied, yet elsewhere advertises file write operations as a core capability. This inconsistency can mislead users into believing the tool is non-persistent when it may actually create or modify local files.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The invocation scope is extremely broad, covering analysis of 'arbitrary input' and many common activities. Overly broad trigger language can cause accidental activation in unrelated contexts, increasing the chance that powerful tools like exec/read/write are used when the user did not intend to invoke this skill.

Vague Triggers

Medium

Confidence: 86% confidence
Finding: The example trigger '执行核心功能' is generic enough to overlap with ordinary user conversation. In an agent environment, such broad phrasing can lead to unintentional invocation and exposure of local tools or side effects without meaningful user intent.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The documentation advertises write capability without a clear warning that local files may be modified or that confirmation should be required. In a tool marketed for analysis, silent or unexpected file modification creates integrity and safety risks for user workspaces.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The skill describes command execution capability without a prominent, user-facing warning about shell risks. Because the skill also accepts broad natural-language tasks, users may not appreciate that a request could translate into actual system command execution with local side effects.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.