Back to skill

Security audit

呼吸感品牌全案

Security checks for vulnerabilities and agentic risk

Overview

The skill provides the advertised brand-design workflow, with ordinary installation risks and no evidence of hidden or harmful behavior.

Review the installation command before running it: it uses changing remote sources, installs globally, and skips confirmation. Prefer reviewing a fixed revision and installing only the intended skill. Image generation may send brand details to your configured image provider. Check the stated noncommercial license before business use.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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 (4)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The supplied command performs software installation rather than the declared brand-design work. Global installation is a material environment change not disclosed in the description. The matching skill name suggests a relationship to the declared purpose, but the installed skill's contents are not supplied, so its design capabilities cannot be verified.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx skills'.

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

Although this is an installation CLI rather than an MCP server, npx skills executes an unpinned npm package, creating a supply-chain risk if a future release or its dependencies are compromised. The command also installs skill content from a repository without an immutable revision, so that content can change independently of this README. This is an optional, user-initiated installation step rather than automatic execution during brand-design work, which reduces exposure; no malicious intent is evident.

Content

No source excerpt is available for this finding.

The skill requires English slogans regardless of the user's chosen output language.

Low
Category
Output Handling
Confidence
98% confidence
Finding

Although the output-language setting follows the user, it explicitly requires brand names and slogans to be in English. The unconditional slogan requirement imposes a specific language without user opt-in; an international branding context does not itself establish that the user wants English copy.

Content

No source excerpt is available for this finding.

The template rules require English brand and slogan text without user opt-in.

Low
Category
Output Handling
Confidence
88% confidence
Finding

Line 11 explicitly requires brand names and slogans to be written in English. Although it separately accommodates Chinese brand names, it does not offer a language choice for slogans or justify the English-only requirement.

Content

No source excerpt is available for this finding.