instagram-dm-sales-agent

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The package is not a hidden backdoor or obfuscated malware, but it is explicitly designed to enable large-scale, covert/anti-detection Instagram cold-outreach (spam) with multi-account/session reuse, headless/cron/Modal deployment, and optional AI calls that leak lead data to an external API — combined with TLS validation disabling and cookie/session injection this makes it a high-risk abusive automation tool that can be used to violate user privacy and platform policies.

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

• Third-party content exposure detected (high risk: 0.90). The skill uses Playwright to navigate and scrape live Instagram pages (e.g., IG_DM_URL = "https://www.instagram.com/direct/inbox/" in scripts like scripts/instagram_dm_sales_agent.py and scripts/ig-dm-browser.js), reads dialog/page content (dialog.textContent(), page.content(), search result text) and acts on those untrusted, user-generated results to select users, verify conversations, and control send/skip logic, so third-party Instagram content can materially influence agent actions.