Added chunking logic to modules and added corresponding regression tests

internal/modules/spoken_word/prompt.go

@@ -24,14 +24,14 @@ type promptTranscriptSection struct {
 
 // BuildProposalMessages mirrors the Python spoken_word-module prompt intent:
 // conservative dysfluency cleanup with strict semantic preservation.
-func BuildProposalMessages(transcript *schema.Transcript, glossary *schema.Glossary) ([]contracts.LLMMessage, error) {
+func BuildProposalMessages(transcript *schema.Transcript, glossary *schema.Glossary, sectionIndex int) ([]contracts.LLMMessage, error) {
 	glossaryJSON, err := json.MarshalIndent(glossary, "", "  ")
 	if err != nil {
 		return nil, fmt.Errorf("marshal glossary prompt context: %w", err)
 	}
 
 	sectionPayload := promptTranscriptSection{
-		SectionIndex: 0,
+		SectionIndex: sectionIndex,
 		Segments:     make([]promptSegment, 0),
 	}
 	if transcript != nil {