marketing-pipeline-share-ai-content

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

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow calls conductResearch() (via app/api/content/generate/route.ts → lib/research/scraper.ts → fetchNewsArticles) which crawls/scrapes external news sources (e.g., TechCrunch/Twitter) and then injects their summary/excerpt text into the LLM prompt as researchData in generateContentWithClaude().

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

• Potentially malicious external URL detected (high risk: 0.90). The skill clearly fetches external research at runtime via RapidAPI (RAPIDAPI_KEY) from sources like TechCrunch / a16z / Twitter and then injects that fetched researchData directly into the LLM prompt (buildPrompt → anthropic / openai calls), so the external API endpoints used for research (via RapidAPI) directly control prompts.