design-inspiration-collector

• Outcome: Pass
• Latest audit: May 28, 2026 at 12:12 PM UTC
• Version:
  1.0.3

Overview

This skill appears to do what it claims: collect design-inspiration links from Dribbble/Pinterest and save local report files, with privacy and file-writing caveats.

Install only if you are comfortable sending design topics to Tavily and saving generated Markdown/JSON reports under ~/design_inspirations. Avoid confidential client names, unreleased product concepts, or regulated data in prompts, and use explicit wording when you want the skill to run.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (7)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 91% confidence
• Finding: The skill uses environment access, network access, and file writing, but does not declare permissions or otherwise make those capabilities explicit. This creates a transparency and policy-enforcement gap: a host may invoke the skill without realizing it can read secrets like API keys, contact external services, and write files, increasing the chance of unintended data exposure or unauthorized side effects.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 95% confidence
• Finding: The documented behavior does not fully match the operational behavior described by the finding: it may write additional JSON output, create directories under the user's home directory, and bypass live Pinterest search by fabricating preset links. This is dangerous because users and the execution environment may consent to a narrower behavior set than what the skill actually performs, leading to unexpected persistence, misleading output provenance, and broader filesystem impact.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The trigger list includes broad, generic phrases such as '找灵感', '设计参考', and platform names like 'Dribbble' and 'Pinterest', which can match normal conversation outside a clear request to invoke the skill. In an agent environment, this can cause unintended activation, leading to unnecessary external searches and file generation without explicit user intent.

Missing User Warnings

Low

• Confidence: 83% confidence
• Finding: The skill is designed to create a Markdown file as a side effect, but the user-facing guidance does not clearly warn that a file will be written. While lower severity, silent file creation can surprise users, clutter workspaces, or be abused in multi-step agent flows where file generation was not expected.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The script persists Markdown and JSON files under the user's home directory automatically, without explicit confirmation, path selection, or disclosure of retention implications. In an agent setting, silent writes can leak sensitive user topics to disk, create unexpected artifacts, and expose data to other local users, backup systems, or later processes that read from the home directory.

Vague Triggers

Medium

• Confidence: 94% confidence
• Finding: The triggers include broad single-word platform names like "Dribbble" and "Pinterest", which can cause the skill to activate on casual mentions rather than explicit requests to collect inspiration. In an agent setting, unintended activation can lead to unnecessary external searches, unwanted file generation, and surprising behavior that overrides a user's actual intent.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The description says the skill should trigger whenever the user needs design references, UI inspiration, or visual ideas, but it does not define clear limits or disambiguation criteria. This increases the chance of over-triggering on general design discussions, causing the agent to take actions like external lookups and document creation when the user may only want advice or brainstorming.

VirusTotal

VirusTotal findings are pending for this skill version.