The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill includes functionality in scripts/vlm_finetune.py and documentation in references/data-formats.md to download image assets from user-provided URLs using the requests library. This is used to convert remote images into the base64 format required for Vision-Language Model (VLM) training.
[COMMAND_EXECUTION]: The provided Python scripts and Markdown documentation utilize the together SDK and CLI to manage model fine-tuning jobs, upload datasets, and create dedicated inference endpoints. These operations are performed against the official Together AI API infrastructure.
[DATA_EXFILTRATION]: The skill facilitates the transfer of local datasets to Together AI's platform (api.together.xyz). This data transfer is the intended core purpose of the skill for model adaptation and training.
[SAFE]: All scripts follow best practices for secret management by retrieving the TOGETHER_API_KEY from environment variables rather than hardcoding credentials. The skill exclusively interacts with well-known AI model providers and infrastructure services.

together-fine-tuning