feat: add remove_prompt(), remove_resource(), and remove_resource_template()

[rgoldstein1989]

Summary

Add removal APIs for prompts, resources, and resource templates to match the existing remove_tool() pattern. This completes the CRUD symmetry for all three MCP primitives, enabling per-instance filtering without reaching into private internals.

• MCPServer.remove_prompt(name) / PromptManager.remove_prompt(name) — raises PromptError
• MCPServer.remove_resource(uri) / ResourceManager.remove_resource(uri) — raises ResourceError
• MCPServer.remove_resource_template(uri_template) / ResourceManager.remove_template(uri_template) — raises ResourceError
• New PromptError exception for symmetry with ToolError and ResourceError

Differences from #2333

This PR covers the same scope as #2333 but adds:

1. Tests — 11 tests mirroring the existing remove_tool test patterns: remove existing, remove nonexistent (error path), remove-and-list via client, and remove-and-access via client for all three primitives. Required to pass CI's 100% coverage gate.
2. AnyUrl | str type on remove_resource — Resources are stored via str(resource.uri) where uri can be AnyUrl. The remove_resource method normalizes with str(uri) before lookup, matching how get_resource() handles it, so callers can pass either type.

Use case

Multi-tenant / multi-instance deployments where the same server image serves different clients. Today users can filter tools per-instance via remove_tool(), but for prompts and resources they're forced to delete from private _manager._prompts / _resources dicts directly.

Fixes #2331

[S0yKaf]

This is useful for my use-case since I can dynamically add prompts to the running server but removing isn't as simple.

As a workaround I do this but it's a bit nasty

mcp._prompt_manager._prompts.pop(name, None)

[maxisbey]

Thanks for the PR, and sorry it sat here without a proper review.

We're closing most of the open PR backlog. v2 is out and changed a lot of the SDK, so many older PRs no longer apply as written, and we're a small team that realistically doesn't have the capacity to work through the rest.

If this still matters to you on v2, the most useful thing you can do is open an issue (or comment on the existing one) with your use case and a repro. Hearing why it matters to you is what we use to decide what to prioritise.

_{AI Disclaimer}