品牌热点营销助手

• Outcome: Review
• Latest audit: May 29, 2026 at 9:29 PM UTC
• Version:
  1.0.0

Overview

This marketing skill is mostly coherent, but it asks users to paste a billable App-Key into chat and documents unclear API handling for that key.

Install only if you trust the Zeelin/Vane service and are comfortable with successful requests spending 100 credits. Prefer a revocable, low-scope key, avoid pasting reusable secrets into chat where transcripts may be retained, and verify the real production endpoint before sending credentials.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Ssd 3

Medium

• Confidence: 97% confidence
• Finding: The skill instructs the agent to ask users for an App-Key directly in chat and then replay that secret in API requests. Collecting reusable credentials in plain conversation increases the chance of accidental disclosure, logging exposure, prompt leakage, and cross-tool misuse; the risk is heightened because the key appears to control billable quota and account access.

Ssd 3

Medium

• Confidence: 96% confidence
• Finding: The example dialogue normalizes a dangerous pattern: users pasting secret keys into chat and the agent operationalizing them. This trains both users and integrators to treat conversational channels as credential entry points, increasing the likelihood of key compromise through transcript retention, model exposure, support access, or downstream prompt injection.

VirusTotal

65/65 vendors flagged this skill as clean.