brand-dna

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow (Step 2 "Fetch Pages" in SKILL.md) uses the WebFetch tool to retrieve arbitrary public website content — homepage, about, and product pages including all visible text, blocks, meta tags and og:image — which the agent then reads and interprets to build brand-profile.json and drive downstream actions, exposing it to untrusted third‑party content that could contain indirect prompt injections.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill uses the WebFetch tool at runtime to retrieve arbitrary external websites and assets (e.g., fetching https://linear.app and linked resources such as https://fonts.googleapis.com/css2?...), and that fetched content is injected and parsed to drive the agent's extraction and output—i.e., remote site content directly controls the agent's prompts/behavior.