fix(agent): skip tools for models without tool-call capability and parse image output

- Add SupportsToolCall to RunConfig; only inject tools into SDK when set
- Update twilight-ai to 497ad09 which adds SSE scanner 10MB buffer
  (fixes token-too-long on large image payloads) and parses the images
  array from OpenAI-compatible chat completions into StreamFilePart

internal/agent/read_media_test.go

@@ -246,6 +246,7 @@ func TestAgentGenerateReadMediaInjectsImageIntoNextStep(t *testing.T) {
 		Model:              &sdk.Model{ID: "mock-model", Provider: modelProvider},
 		Messages:           []sdk.Message{sdk.UserMessage("look at the image")},
 		SupportsImageInput: true,
+		SupportsToolCall:   true,
 		Identity: SessionContext{
 			BotID: "bot-1",
 		},
@@ -321,6 +322,7 @@ func TestAgentGenerateReadMediaInjectsAnthropicSafeImageIntoNextStep(t *testing.
 		Model:              &sdk.Model{ID: "mock-model", Provider: modelProvider},
 		Messages:           []sdk.Message{sdk.UserMessage("look at the image")},
 		SupportsImageInput: true,
+		SupportsToolCall:   true,
 		Identity: SessionContext{
 			BotID: "bot-1",
 		},
@@ -367,6 +369,7 @@ func TestAgentStreamReadMediaPersistsInjectedImageInTerminalMessages(t *testing.
 		Model:              &sdk.Model{ID: "mock-model", Provider: modelProvider},
 		Messages:           []sdk.Message{sdk.UserMessage("look at the image")},
 		SupportsImageInput: true,
+		SupportsToolCall:   true,
 		Identity: SessionContext{
 			BotID: "bot-1",
 		},