modelscope-skills / agents

You have additional SKILLs documented in directories containing a "SKILL.md" file.

These skills are:

• gradio -> "skills/modelscope-gradio/SKILL.md"
• modelscope-cli -> "skills/modelscope-cli/SKILL.md"
• modelscope-dataset-explorer -> "skills/modelscope-dataset-explorer/SKILL.md"
• modelscope-datasets -> "skills/modelscope-datasets/SKILL.md"
• modelscope-evaluation -> "skills/modelscope-evaluation/SKILL.md"
• modelscope-jobs -> "skills/modelscope-jobs/SKILL.md"
• modelscope-local-trainer -> "skills/modelscope-local-trainer/SKILL.md"
• modelscope-local-vision-trainer -> "skills/modelscope-local-vision-trainer/SKILL.md"
• modelscope-mcp -> "skills/modelscope-mcp/SKILL.md"
• modelscope-model-trainer -> "skills/modelscope-model-trainer/SKILL.md"
• modelscope-paper-publisher -> "skills/modelscope-paper-publisher/SKILL.md"
• modelscope-tool-builder -> "skills/modelscope-tool-builder/SKILL.md"
• modelscope-tracking -> "skills/modelscope-tracking/SKILL.md"
• modelscope-vision-trainer -> "skills/modelscope-vision-trainer/SKILL.md"
• transformers-js -> "skills/transformers.js/SKILL.md"

IMPORTANT: You MUST read the SKILL.md file whenever the description of the skills matches the user intent, or may help accomplish their task.

<available_skills>

gradio: Build or edit Gradio apps from plain-language UI requests, choosing the simplest valid Gradio pattern for the job. modelscope-cli: Execute ModelScope Hub login, create, download, and upload tasks from plain-language requests with the official CLI. modelscope-dataset-explorer: Explore ModelScope datasets from plain-language inspection requests using metadata lookup, selective download, and local profiling. modelscope-datasets: Build, clean, validate, and publish ModelScope datasets from plain-language instructions with automatic template selection. modelscope-evaluation: Turn plain-language evaluation requests into benchmark extraction, remote evaluation jobs, and publishable result artifacts. modelscope-jobs: Execute plain-language batch-job requests as reproducible local or PAI DLC workflows and publish the outputs to ModelScope. modelscope-local-trainer: Turn plain-language local-training requests into editable configs and launch scripts, then run them after the user fills a local config file and replies continue. modelscope-local-vision-trainer: Turn plain-language local vision-training requests into editable configs and launch scripts, then run them after the user fills a local config file and replies continue. modelscope-mcp: Configure the official ModelScope MCP server from plain-language setup requests and use it for tool-based Hub discovery. modelscope-model-trainer: Turn plain-language language-model training requests into ms-swift configs, PAI DLC jobs, and ModelScope publications with sensible defaults. modelscope-paper-publisher: Turn plain-language publication requests into evidence-backed Markdown reports for ModelScope repositories. modelscope-tool-builder: Convert plain-language ModelScope automation needs into reusable scripts with stable flags, structured output, and official API usage. modelscope-tracking: Add local, file-based experiment tracking to ModelScope workflows from plain-language requests. modelscope-vision-trainer: Turn plain-language vision fine-tuning requests into task plans, config files, and PAI-ready submission assets. transformers-js: Build browser or Node.js inference workflows from plain-language requests using Transformers.js and locally downloaded model assets. </available_skills>

Paths referenced within SKILL folders are relative to that SKILL. For example, my-skill/scripts/example.py is referenced as my-skill/scripts/example.py.