blog-cover-generator

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The prompt requires running CLI commands that include API keys/Client IDs as literal arguments (e.g., config set-key <YOUR_GEMINI_API_KEY>), which encourages asking for or embedding secrets verbatim in agent output even though the CLI stores them securely — this creates an exfiltration risk.

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

• Third-party content exposure detected (high risk: 0.75). The runtime workflow uses the -l/--logo domain to fetch and process third-party logos via Brandfetch (outsider-authored web content), and then injects that fetched logo into the model payload for Gemini image generation.