google-ad-scraper

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The script intentionally sends API tokens/requests through an external proxy (gooseworks.ai) by reading a required GOOSEWORKS_API_KEY env var and including it in HTTP requests, which can leak sensitive credentials and constitutes potential data exfiltration to a third-party service.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). Yes—during advertiser ID resolution the script calls Apify’s apify~web-scraper to load Google Ads Transparency Center pages at runtime and then ingests the extracted page text/links (outsider-authored web content) from the dataset items into the program’s output/LLM context.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The script posts user-supplied JavaScript (pageFunction) and starts Apify actors via the Gooseworks proxy at https://app.gooseworks.ai (BASE_URL /v1/proxy/apify), which causes remote code execution on that external service.