auth-manager

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly instructs reading tokens/cookies from local session files and embedding them into curl requests (Authorization/Cookie headers), which requires including secret values verbatim in commands/requests and is high-risk.

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly navigates to and snapshots public third‑party sites (e.g., browser(action='navigate', targetUrl='https://polymarket.com' / 'https://x.com/home' etc.) and uses those page snapshots and configured "logged_in_indicators" from ~/.openclaw/auth-platforms.json to determine status, so untrusted/user-generated web content is ingested and can influence agent decisions and actions.