Complete MCP tool expansion (36 → 53 tools)

[kiyeonjeon21]

Goal

Expand MCP server from 36 to 53 tools. Currently CLI has ~69 commands but MCP only exposes 36 (52%).

Remaining work

Phase 2c: Image tools (+2)

• generate_image — extract from cli/src/commands/ai-image.ts
• edit_image — extract from cli/src/commands/ai-image.ts

Phase 2d: Video tools (+4)

• generate_video — extract from cli/src/commands/ai-video.ts (4 providers)
• generate_video_status — polling endpoint
• generate_video_cancel — cancel generation
• generate_video_extend — Kling/Veo extension

Phase 3: Audio tools (+6)

• audio_transcribe — Whisper transcription
• audio_voice_clone — ElevenLabs voice cloning
• audio_dub — multilingual dubbing
• audio_isolate — vocal extraction
• audio_duck — auto-ducking
• audio_restore — noise reduction

Phase 4: Remaining (+5)

• generate_storyboard, generate_background, edit_grade, edit_reframe, edit_speed_ramp

Pattern

Each tool follows the same extraction pattern:

1. Define Options/Result interfaces in CLI command file
2. Extract .action() logic into export async function executeXxx()
3. Remove ora/chalk/console.log → return structured result
4. Add MCP wrapper in mcp-server/src/tools/
5. Register in mcp-server/src/tools/index.ts

Reference

See plan document for detailed file paths and line numbers.