Security audit

Logo设计指南免费版

Security checks across malware telemetry and agentic risk

Overview

This is mostly a logo-design guide, but it requests command and file-writing authority that is not clearly needed or bounded.

Review before installing. The content does not show malicious code, exfiltration, destructive commands, or persistence, but a logo-design guide should not normally need command execution or file-writing access. Prefer a version that removes exec/write and clearly documents or removes callback_url behavior.

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 (6)

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: The skill declares `exec` and `write` capabilities even though the content is a static logo-design guide with no legitimate need to modify files or run commands. This increases attack surface by allowing a low-risk reference skill to be invoked with unnecessary privileges, which could enable unintended filesystem changes or command execution if the surrounding agent trusts the manifest.

Description-Behavior Mismatch

Low

Confidence: 80% confidence
Finding: Documenting `callback_url` and asynchronous-style handling suggests the skill may transmit data to external endpoints, yet no privacy, validation, or network-safety constraints are provided. In an agent ecosystem, this can normalize unsafe user input collection or cause operators to assume network behavior exists and is safe when it is undocumented and unjustified for a static design guide.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: Overly broad scope language can cause the agent to invoke this skill in situations beyond logo-design guidance, especially because the skill also advertises elevated capabilities. Misrouting a broadly-triggered skill with `exec` and `write` permissions increases the chance of unintended actions in unrelated workflows.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The coverage keywords are fragmented and ambiguous, which can trigger accidental invocation based on partial phrase matching instead of clear user intent. In context, this is more dangerous because the skill is not just informational on paper—it also requests command and file-write capabilities, magnifying the consequences of false invocation.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: Advertising `write` and `exec` without any user-facing warning about system modification risks violates least-astonishment and can lead users or orchestrators to authorize a seemingly harmless design guide with dangerous privileges. The mismatch between benign description and elevated capability makes this more concerning than if it were a clearly operational/admin skill.

Missing User Warnings

Medium

Confidence: 82% confidence
Finding: Documenting callback URL support without warning about external transmission, privacy exposure, or destination trust can mislead users into supplying endpoints and data without understanding the risks. In agent environments, this can facilitate data exfiltration, SSRF-like misuse, or privacy violations if later implemented naively.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.