amazon-ppc-campaign

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and parses open/public third-party pages (notably via scripts/fetch-competitor.sh which curls Amazon product pages and via web_search/autocomplete calls) and then ingests titles, bullets, search terms, and competitor data as core inputs that directly drive campaign-building decisions, so untrusted external content can influence agent actions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill's runtime calls to external Amazon endpoints (e.g., fetching product pages like https://www.amazon.com/dp/ via scripts/fetch-competitor.sh and the autocomplete API https://completion.amazon.com/api/2017/suggestions?mid=ATVPDKIKX0DER&alias=aps&prefix=...) are performed at runtime and the returned content is parsed and used to generate keywords and campaign instructions that directly control the agent's outputs.