Skip to main content
Beyond its tools, Cotality MCP exposes reusable prompt templates and reference resources for the property data tools. They cost nothing extra to use and encode the calling patterns and domain conventions you would otherwise have to write yourself.

Discovering prompts and resources

Both follow the same request pattern as tools/list.
Like tools/list, these calls count against your rate limit. Call them once when the session opens rather than before every tool call.

Prompts

Arguments shown with a default value are optional. All other arguments are required.
MCP prompt arguments are passed as strings. Where an argument is conceptually a number or a list (clips), send it as a string or a JSON scalar — the server coerces it. This is different from tool arguments, which are validated against a strict JSON schema.

CLIP lookup

Property characteristics

Climate and risk

Prompt names are shown as the servers register them. Unlike tool names, they do not carry a server prefix today, and the two CLIP lookup prompts use hyphens while the rest use underscores. Always use the exact name returned by prompts/list rather than composing one.

Resources

Address a resource by its URI — that is what resources/read takes. The Name column is the display label the server registers; the resource://clip-mcp/info resource registers no explicit label, so it surfaces under its function name.

Scopes

Scope enforcement applies to tools only. Prompts and resources are not scope-filtered, so prompts/list and resources/list can return entries for a server whose tools you cannot call.
A prompt being listed does not mean you are entitled to the tools it tells the model to call. Confirm entitlement with tools/list before you build a flow around a prompt.