feat(ui): Video playground tab + video capability plumbing

New Video tab in the Playground: model selector filtered to
video_generation/image_to_video capabilities, prompt + negative prompt,
size/frames/fps/steps/guidance/seed knobs (0 = model default), optional
conditioning-image upload (image-to-video), elapsed-time spinner for
the minutes-long blocking call, inline <video> playback + download.
Calls POST /v1/videos/sync (multipart); resolution rides as both
width/height and size so either upstream convention honors it.

Config side: "video" joins the valid capability modalities, and
/v1/models maps text->video to video_generation and image->video to
image_to_video, mirroring the image mappings. Declare e.g.:

  videogen-wan22-5b:
    capabilities:
      in: [text, image]
      out: [video]

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AXQxVhXBw8PwFAtsVrXSmj

internal/config/model_config.go

@@ -14,6 +14,7 @@ var validModalities = map[string]struct{}{
 	"text":  {},
 	"audio": {},
 	"image": {},
+	"video": {},
 }
 
 // ModelCapConfig defines what modalities and features a model supports.
@@ -37,12 +38,12 @@ func (c ModelCapConfig) Empty() bool {
 func (c ModelCapConfig) Validate() error {
 	for _, m := range c.In {
 		if _, ok := validModalities[m]; !ok {
-			return fmt.Errorf("capabilities.in: invalid modality %q, must be one of: text, audio, image", m)
+			return fmt.Errorf("capabilities.in: invalid modality %q, must be one of: text, audio, image, video", m)
 		}
 	}
 	for _, m := range c.Out {
 		if _, ok := validModalities[m]; !ok {
-			return fmt.Errorf("capabilities.out: invalid modality %q, must be one of: text, audio, image", m)
+			return fmt.Errorf("capabilities.out: invalid modality %q, must be one of: text, audio, image, video", m)
 		}
 	}
 	if c.Context < 0 {