Back to plugin

Security audit

Creatify Ad Agent

Security checks for vulnerabilities and agentic risk

Overview

The package provides a coherent video-ad workflow with disclosed cloud processing, credit costs, and saved projects.

Give an explicit advertising request and set a credit budget. Only supply files and voice recordings approved for processing by Creatify and ElevenLabs, and expect project assets to remain stored remotely.

Vulnerability Patterns
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Pasting any store or product URL can activate the skill without an advertising request.

Medium
Category
Trigger Abuse
Confidence
97% confidence
Finding

The activation description treats a pasted store or product URL as sufficient authorization to start an advertising workflow, even when the user might want something else. The workflow creates an external project and imports page assets before requiring plan confirmation; the paid-generation safeguard limits financial exposure but does not prevent unintended external processing.

Content

No source excerpt is available for this finding.

The transcription workflow sends audio to a third-party service without a privacy warning.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The workflow mandates transcription through ElevenLabs Scribe, including for user-provided dialogue, without requiring disclosure or authorization before sending audio to that external processor. This could expose confidential campaign material, personal information, or identifiable voice recordings outside the user's expected Creatify processing boundary. Transcription is legitimate for ad production and the provider is named, which reduces suspicion of malicious exfiltration but does not establish user consent or appropriate data handling.

Content

No source excerpt is available for this finding.

The workflow uploads local attachments to an external service without a user-facing privacy warning.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The skill instructs the agent to upload locally attached files and import their CDN URLs into the hosted project. Although it identifies Creatify as the hosted provider, it does not instruct the agent to notify users that their local files will be transferred to an external service.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.