Minimax-Multimodal-Toolkit

**Switch to `mmx-cli`: major simplification and platform change** - Replaces all separate bash scripts and API reference docs with a unified command-line interface (`mmx`). - Removes 15 legacy shell script and markdown reference files. - Updates skill name and description to reflect the new `mmx-cli` interface. - New workflow: install `mmx-cli` via npm and use a single CLI for text, image, video, speech, and music generation. - All instructions, options, and usage details are consolidated into the new SKILL.md focused on `mmx` commands and agent-friendly flags.

**MiniMax multimodal toolkit now uses bash scripts, adds image generation, and drops Python dependencies:** - Switched all major scripts from Python to bash, requiring only `ffmpeg`, `jq`, and `curl` (no pip/Python needed) - Added support for image generation (text-to-image, image-to-image) via new image generation scripts - All scripts and workflows updated to use `.sh` bash scripts instead of `.py` Python scripts - Requires new environment variable `MINIMAX_API_HOST` (in addition to API key) with region selection instructions - Updated and simplified prerequisites and usage instructions; Python environment setup is no longer needed

MiniMax multimodal skill initial release — generate speech, music, video, and process media via MiniMax AI. - Supports TTS (text-to-speech), voice cloning, and custom voice design. - Music and video creation: text-to-video, image-to-video, templates, multi-scene, and more. - FFmpeg-powered media tools: convert, concatenate, trim, extract audio/video. - Enforces all outputs to agent’s minimax-output/ directory, with explicit output paths required. - Includes comprehensive rules for TTS voice segmentation (single/multi-voice) and API key setup guidance.