togethercomputer/skills

togethercomputer/skills

Description: Skills to help your coding agents use Together AI products.

Language: Python

License: MIT

Stars: 32

Forks: 4

Open issues: 4

Created: 2026-03-31T18:07:13Z

Pushed: 2026-06-10T16:58:36Z

Default branch: main

Fork: no

Archived: no

README:

Together AI Skills for Coding Agents

A collection of 12 agent skills that provide comprehensive knowledge of the Together AI platform — inference, training, embeddings, audio, video, images, function calling, and infrastructure.

Each skill teaches AI coding agents how to use a specific Together AI product, including API patterns, SDK usage (Python and TypeScript), CLI commands, direct API usage, model selection, and best practices. Skills include runnable Python scripts (using the Together Python v2 SDK), TypeScript examples, and CLI/API workflow guidance.

Compatible with Claude Code, Cursor, Codex, and Gemini CLI.

What Are Skills?

Skills are markdown instruction files that give AI coding agents domain-specific knowledge. When an agent detects that a skill is relevant to your task, it loads the skill's instructions and uses them to write better code.

Each skill contains:

• `SKILL.md` — Lean routing guidance for the agent: when to use the skill, when to hand off, and where to look next
• `references/` — Detailed reference docs (model lists, API parameters, CLI commands)
• `scripts/` — Runnable Python scripts demonstrating complete workflows
• `agents/openai.yaml` — Optional UI metadata for OpenAI/Codex surfaces

Skills Overview

| Skill | Description | Scripts | |-------|-------------|---------| | together-chat-completions | Real-time and streaming text generation via Together AI's OpenAI-compatible chat/completions API, including multi-tur... | async_parallel.py, chat_basic.py, debug_headers.py, reasoning_models.py, structured_outputs.py, tool_call_loop.py | | together-images | Text-to-image generation and image editing via Together AI, including FLUX and Kontext models, LoRA-based styling, re... | generate_image.py, kontext_editing.py, lora_generation.py | | together-video | Text-to-video and image-to-video generation via Together AI, including keyframe control, model and dimension selectio... | generate_video.py, image_to_video.py | | together-audio | Text-to-speech and speech-to-text via Together AI, including REST, streaming, and realtime WebSocket TTS, plus transc... | stt_realtime.py, stt_transcribe.py, tts_generate.py, tts_websocket.py | | together-embeddings | Dense vector embeddings, semantic search, RAG pipelines, and reranking via Together AI. | embed_and_rerank.py, rag_pipeline.py, semantic_search.py | | together-fine-tuning | LoRA, full fine-tuning, DPO preference tuning, VLM training, function-calling tuning, reasoning tuning, and BYOM uplo... | dpo_workflow.py, finetune_workflow.py, function_calling_finetune.py, reasoning_finetune.py, vlm_finetune.py | | together-batch-inference | High-volume, asynchronous offline inference at up to 50% lower cost via Together AI's Batch API. | batch_workflow.py | | together-evaluations | LLM-as-a-judge evaluation framework on Together AI. | run_evaluation.py | | together-sandboxes | Remote Python execution in managed sandboxes on Together AI with stateful sessions, file uploads, data analysis, char... | execute_with_session.py | | together-dedicated-endpoints | Single-tenant GPU endpoints on Together AI with autoscaling and no rate limits. | deploy_finetuned.py, manage_endpoint.py, upload_custom_model.py | | together-dedicated-containers | Custom Dockerized inference workers on Together AI's managed GPU infrastructure. | queue_client.py, sprocket_hello_world.py | | together-gpu-clusters | On-demand and reserved GPU clusters (H100, H200, B200) on Together AI with Kubernetes or Slurm orchestration, shared ... | manage_cluster.py, manage_storage.py |

Installation

Quick Install (Any Agent)

Install all skills at once using skills.sh:

npx skills add togethercomputer/skills

This works with Claude Code, Cursor, Codex, and other agents that support the Agent Skills specification.

Claude Code

cp -r skills/together-* your-project/.claude/skills/
# Global availability
cp -r skills/together-* ~/.claude/skills/

Marketplace plugin coming soon.

Cursor

cp -r skills/together-* your-project/.cursor/skills/

Cursor plugin marketplace listing coming soon.

Codex

cp -r skills/together-* your-project/.agents/skills/

Gemini CLI

gemini extensions install https://github.com/togethercomputer/skills.git --consent

Verify installation

# Claude Code
ls your-project/.claude/skills/together-*/SKILL.md
# Codex
ls your-project/.agents/skills/together-*/SKILL.md

You should see one SKILL.md per installed skill.

Usage

Once installed, skills activate automatically when the agent detects a relevant task. No explicit invocation is needed.

Examples

Chat completions — Ask the agent to build a chat app:

> Build a multi-turn chatbot using Together AI with Llama 3.3 70B

The agent will use the together-chat-completions skill to generate correct v2 SDK code with proper model IDs, parameters, and streaming patterns.

Function calling — Ask for tool-using agents:

> Create an agent that can check weather and stock prices using Together AI function calling

The agent will reference together-chat-completions for the complete tool call loop pattern, including parallel tool calls and tool_choice options.

Image generation — Ask for image workflows:

> Generate a FLUX image with Together AI and save it locally as PNG

The agent will use together-images to write code with the correct model ID, base64 decoding, and file saving.

Fine-tuning — Ask to fine-tune a model:

> Fine-tune Llama 3.1 8B on my dataset using Together AI with LoRA

The agent will reference together-fine-tuning for data format requirements, training parameters, monitoring, and deployment.

Using the scripts

Each script is a standalone, runnable example. They require the Together Python SDK and an API key:

uv pip install "together>=2.0.0"
export…