ad-creative

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted performance data from users or APIs to analyze and generate new ad variations. * Ingestion points: Ad performance data (CSV, paste, or API output) as described in SKILL.md and evals.json. * Boundary markers: None identified in the instructions for data processing. * Capability inventory: Integration with google-ads, meta-ads, linkedin-ads, and tiktok-ads CLI tools for campaign and report management. * Sanitization: No sanitization or validation of the ingested performance data is mentioned.
  • [COMMAND_EXECUTION]: The skill provides instructions and examples for executing local shell commands and CLI tools to interact with advertising platforms, such as running 'node tools/clis/google-ads.js'.
  • [EXTERNAL_DOWNLOADS]: The skill references numerous external AI service APIs (Google Gemini, OpenAI, ElevenLabs, etc.) and provides instructions for cloning and setting up external repositories such as Voicebox from GitHub.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 11:22 PM
Security Audit — agent-trust-hub — ad-creative