)

Add support for prompts and resources to the mcptest server, including:
- New methods to add prompts and resources to test servers
- Batch methods for adding multiple prompts/resources at once
- Test cases demonstrating prompt and resource functionality
- Helper methods for resource capabilities registration

…339)

* feat(mcptest): Change `Server.Start` to accept a `context.Context`.

Previously, `mcptest` used `context.TODO()` as the context with this server
code was executed (with a comment to upgrade to `testing.T.Context()`
eventually.

This effectively prevents tests from preparing a context to be used with server
code. This is imporant, because `WithHTTPContextFunc` and friends appear to be
the intended way to extract things like authentication information from the raw
HTTP request and pass it to the server code. Without this change, such code is
effectively untestable, at least with this package.

The `NewServer` convenience method has been left unchanged (i.e. it does not
accept a context) to avoid breaking users taking the happy path. Chaning the
signature of `Start()` is however a breaking change.

* test(mcptest): Update tests to use the new `Start()` signature.

---------

Co-authored-by: Navendu Pottekkat <navendu@apache.org>

* Scaffold documentation site

* Remove node_modules

* rename

* add getting-started

* add gh-pages workflow

* fix

* Add more docs

* corrections

* Fix mistakes

* fix mistakes

* cleanup

* fixes

* Apply suggestions from code review

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

* Update www/docs/pages/clients/transports.mdx

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

* Update www/docs/pages/clients/transports.mdx

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

* Update www/docs/pages/clients/transports.mdx

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

* Update www/docs/pages/transports/inprocess.mdx

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

* fixes

* fixes

* fixes

* Update www/docs/pages/servers/prompts.mdx

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

* Update www/docs/pages/transports/stdio.mdx

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

---------

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

…365)

* fix: SSE parser now correctly handles events without event field

According to the W3C SSE specification, the event field is optional.
SSE events with only data fields should be processed normally and
treated as "message" events (the default event type).

This fixes issue #369 where the SSE parser incorrectly required both
event and data fields, causing "unexpected nil response" errors when
servers sent valid SSE events with only data fields.

🤖 Generated with opencode
Co-Authored-By: opencode <noreply@opencode.ai>

* fix: resolve race condition in SSE test

Fixed race condition in TestSSE/SSEEventWithoutEventField where multiple
goroutines were accessing the same HTTP response writer concurrently.
Simplified the test to use channel-based synchronization instead of
shared state between HTTP handlers.

🤖 Generated with opencode
Co-Authored-By: opencode <noreply@opencode.ai>

* fix: StreamableHTTP SSE parser now correctly handles events without event field

- Fixed SSE parser in StreamableHTTP transport to process events with only data field
- According to W3C SSE specification, the event field is optional
- When no event type is specified, defaults to "message" event type
- Added comprehensive test case to verify the fix works correctly
- Resolves issue #369 where SSE events without event field caused "unexpected nil response" errors

🤖 Generated with opencode
Co-Authored-By: opencode <noreply@opencode.ai>

---------

Co-authored-by: opencode <noreply@opencode.ai>

…375)

Optimize docs