mobbin

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

• Third-party content exposure detected (medium risk: 0.30). SKILL.md mandates using Mobbin MCP tools (search_screens, search_flows, search_sections) to fetch and read returned images/metadata/links, so outsider-supplied free-text can be used to craft the MCP natural-language query that the runtime ingests via the MCP response.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill mandates runtime calls to the Mobbin MCP server at https://api.mobbin.com/mcp and requires using its returned references to drive design outputs, so remote content directly controls the agent's prompts/results.