twitter-GTM-find-Skill

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). These are Twitter links and third‑party GitHub repositories (Xquik-dev); although not direct executable downloads, installing or running plugins/code from an unvetted GitHub account can deliver executable code and therefore poses a moderate-to-high risk unless you review the repo, check its popularity/commit history, and verify the maintainer.

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

• Third-party content exposure detected (high risk: 0.85). The runtime path scrapeTwitterJobs() → index.ts ingests itemAny.full_text and itemAny.user_info.description from Apify’s Twitter dataset (outsider-authored tweets/bios), then passes that free text into the LLM via extractJobDetails(prompt) and later into evaluateICP(prompt) for Google-search grounded validation.

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

• Potentially malicious external URL detected (high risk: 0.85). The skill invokes Google Generative AI (Gemini) at runtime via the @google/generative-ai SDK (generativeai.googleapis.com), which generates JSON outputs that directly control extraction and ICP decisions, and it also calls a remote Apify actor (ghSpYIW3L1RvT57NT) via the Apify API (api.apify.com) to execute scraping code at runtime — both are runtime external dependencies that influence agent behavior.