explore-recipes

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

• Potentially malicious external URL detected (high risk: 0.80). The skill requires the AI tool to connect to the MCP endpoint https://api.shipswift.app/mcp and uses runtime calls like listRecipes/getRecipe to fetch full recipe implementations (external content injected into the agent's context), so remote content can directly supply code/instructions used by the agent.