design-library

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

• Third-party content exposure detected (high risk: 0.85). The skill fetches third-party DESIGN.md from the public GitHub raw URL (https://raw.githubusercontent.com/VoltAgent/awesome-design-md/.../DESIGN.md) and then hands the fetched markdown text to design-md, so outsider-authored free-form prose/tokens from that file can enter the agent’s LLM context at runtime.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches the raw DESIGN.md at runtime from https://raw.githubusercontent.com/VoltAgent/awesome-design-md/main/design-md//DESIGN.md and then uses that file as the authoritative project taste layer (a prompt/data input), making it a high-confidence prompt-injection surface that can directly control agent behavior.