### Quick Start Examples

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/research/summarize/SKILL.md

Demonstrates basic usage for summarizing URLs, local files, and YouTube links. Ensure the 'summarize' binary is installed and accessible.

```bash
summarize "https://example.com" --model google/gemini-3-flash-preview
```

```bash
summarize "/path/to/file.pdf" --model google/gemini-3-flash-preview
```

```bash
summarize "https://youtu.be/dQw4w9WgXcQ" --youtube auto
```

--------------------------------

### Install and Start Coqui TTS Server (uv)

Source: https://github.com/moltis-org/moltis/blob/main/docs/src/voice.md

Install Coqui TTS with server capabilities using uv and start the TTS server. Ensure torch and related libraries are installed with the correct backend.

```bash
# Via uv
uv pip install torch torchaudio torchcodec --torch-backend=auto
uv pip install 'coqui-tts[server]'
tts-server --model_name tts_models/en/ljspeech/tacotron2-DDC
```

--------------------------------

### Install Ollama and Serve Model

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/unsloth/references/llms-txt.md

Installs Ollama and starts the Ollama server in the background. It then runs a specific DeepSeek model using Ollama.

```bash
apt-get update
apt-get install pciutils -y
curl -fsSL https://ollama.com/install.sh | sh
```

```bash
OLLAMA_MODELS=unsloth ollama serve &

OLLAMA_MODELS=unsloth ollama run hf.co/unsloth/DeepSeek-V3.1-Terminus-GGUF:TQ1_0
```

--------------------------------

### Install Dependencies and Run Local Development Server

Source: https://github.com/moltis-org/moltis/blob/main/docs/README.md

Navigate to the docs directory, install dependencies using npm ci, and start the local development server with npm run dev.

```bash
cd docs
npm ci
npm run dev
```

--------------------------------

### Install Ollama and Serve Models

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/unsloth/references/llms-txt.md

Installs Ollama and starts a server, setting the OLLAMA_MODELS environment variable to 'unsloth'. This is useful for managing Unsloth models within Ollama.

```bash
apt-get update
apt-get install pciutils -y
curl -fsSL https://ollama.com/install.sh | sh
```

```bash
OLLAMA_MODELS=unsloth ollama serve &

OLLAMA_MODELS=unsloth ollama run hf.co/unsloth/GLM-4.6-GGUF:TQ1_0
```

--------------------------------

### Install and Verify Audiocraft

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/models/audiocraft-audio-generation/references/troubleshooting.md

Install the audiocraft library from PyPI or GitHub and verify the installation.

```bash
pip install audiocraft
```

```bash
pip install git+https://github.com/facebookresearch/audiocraft.git
```

```python
python -c "from audiocraft.models import MusicGen; print('OK')"
```

--------------------------------

### Basic SFT Training Setup

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/fine-tuning-with-trl/references/sft-training.md

Sets up and initiates a basic Supervised Fine-Tuning (SFT) process using the TRL library. This involves loading a model, tokenizer, dataset, configuring training parameters, and starting the training loop.

```python
from trl import SFTTrainer, SFTConfig
from transformers import AutoModelForCausalLM, AutoTokenizer
from datasets import load_dataset

# Load model
model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-0.5B")
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-0.5B")

# Load dataset
dataset = load_dataset("trl-lib/Capybara", split="train")

# Configure
config = SFTConfig(
    output_dir="Qwen2.5-SFT",
    per_device_train_batch_size=4,
    num_train_epochs=1,
    learning_rate=2e-5,
    save_strategy="epoch"
)

# Train
trainer = SFTTrainer(
    model=model,
    args=config,
    train_dataset=dataset,
    tokenizer=tokenizer
)
trainer.train()
```

--------------------------------

### Example Usage Placeholder

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/github/github-issues/templates/feature-request.md

This is a placeholder for example usage. Replace with actual command-line examples or API sketches.

```shell
# Example usage
```

--------------------------------

### Install Moltis with cURL

Source: https://github.com/moltis-org/moltis/blob/main/website/index.en.html

Installs Moltis using a shell script. This is a quick way to get started.

```bash
$ curl -fsSL https://www.moltis.org/install.sh | sh
```

--------------------------------

### Install uv and Set Up Environment

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/unsloth/references/llms-txt.md

Download and install uv, then set up a virtual environment with a specified Python version.

```bash
curl -LsSf https://astral.sh/uv/install.sh | sh && source $HOME/.local/bin/env
mkdir 'unsloth-blackwell' && cd 'unsloth-blackwell'
uv venv .venv --python=3.12 --seed
```

--------------------------------

### Bash Commands for Obsidian Headless Setup

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/research/llm-wiki/SKILL.md

These bash commands guide the setup of `obsidian-headless` for servers and headless machines. It includes installation, login, remote vault creation, and continuous synchronization.

```bash
# Requires Node.js 22+
npm install -g obsidian-headless

# Login (requires Obsidian account with Sync subscription)
ob login --email <email> --password '<password>'

# Create a remote vault for the wiki
ob sync-create-remote --name "LLM Wiki"

# Connect the wiki directory to the vault
cd ~/wiki
ob sync-setup --vault "<vault-id>"

# Initial sync
ob sync

# Continuous sync (foreground — use systemd for background)
ob sync --continuous

```

--------------------------------

### One-Command Gateway Setup Wizard

Source: https://github.com/moltis-org/moltis/blob/main/plans/2026-04-20-plan-self-improving-agent-loop.md

This bash command sequence demonstrates the user onboarding process for setting up the Moltis agent. It involves fetching the installer, running the setup wizard, and starting the agent.

```bash
curl -sSL https://install.moltis.org | sh
moltis setup
moltis serve
```

--------------------------------

### cli.main.fetch

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/axolotl/references/api.md

Fetches example configs or other resources.

```APIDOC
## fetch

### Description
Fetches example configurations or other resources from the Axolotl repository.

### Parameters
- directory: The directory to fetch resources from (e.g., 'examples', 'deepspeed_configs').
- dest: The destination path to save the fetched resources.
```

--------------------------------

### Setup and Basic Presentation Structure

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/productivity/powerpoint/pptxgenjs.md

Initialize a new presentation, set layout and author details, add a slide with text, and save the presentation to a file.

```javascript
const pptxgen = require("pptxgenjs");

let pres = new pptxgen();
pres.layout = 'LAYOUT_16x9';  // or 'LAYOUT_16x10', 'LAYOUT_4x3', 'LAYOUT_WIDE'
pres.author = 'Your Name';
pres.title = 'Presentation Title';

let slide = pres.addSlide();
slide.addText("Hello World!", { x: 0.5, y: 0.5, fontSize: 36, color: "363636" });

pres.writeFile({ fileName: "Presentation.pptx" });
```

--------------------------------

### Start Moltis

Source: https://github.com/moltis-org/moltis/blob/main/docs/src/installation.md

Command to start the Moltis application after installation.

```bash
moltis
```

--------------------------------

### Install uv and Set Up Project Environment

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/unsloth/references/llms-full.md

Install the `uv` tool and create a new project directory with a virtual environment configured for Python 3.12.

```bash
curl -LsSf https://astral.sh/uv/install.sh | sh && source $HOME/.local/bin/env
mkdir 'unsloth-blackwell' && cd 'unsloth-blackwell'
uv venv .venv --python=3.12 --seed
source .venv/bin/activate
```

--------------------------------

### Install Axolotl with Dependencies

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/axolotl/references/other.md

Installs Axolotl along with specific dependencies like flash-attn and deepspeed. It also fetches example configurations.

```bash
pip3 install -U packaging==23.2 setuptools==75.8.0 wheel ninja
pip3 install --no-build-isolation axolotl[flash-attn,deepspeed]

# Download example axolotl configs, deepspeed configs
axolotl fetch examples
axolotl fetch deepspeed_configs  # OPTIONAL
```

--------------------------------

### p5.js 1.x vs 2.x: Async Setup

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/creative/p5js/references/core-api.md

Illustrates the change from `preload()` to `async setup()` in p5.js 2.0 for loading assets. This allows for more flexible asynchronous operations during initialization.

```javascript
// p5.js 1.x
let img;
function preload() { img = loadImage('cat.jpg'); }
function setup() { createCanvas(800, 800); }

// p5.js 2.x
let img;
async function setup() {
  createCanvas(800, 800);
  img = await loadImage('cat.jpg');
}
```

--------------------------------

### Install Mod Loader with Env Var

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/gaming/minecraft-modpack-server/SKILL.md

Install the mod loader (NeoForge/Forge) without launching the server by using the `INSTALL_ONLY` environment variable with the start script, or by running the Forge installer directly.

```bash
cd ~/minecraft-server/server
ATM10_INSTALL_ONLY=true bash startserver.sh
# Or for generic Forge packs:
# java -jar forge-*-installer.jar --installServer
```

--------------------------------

### Start llama-server with basic and GPU options

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/inference/llama-cpp/references/server.md

Use these commands to start the llama-server with or without GPU acceleration. Specify the model path and other configurations.

```bash
# Basic server
./llama-server \
    -m models/llama-2-7b-chat.Q4_K_M.gguf \
    --host 0.0.0.0 \
    --port 8080 \
    -c 4096  # Context size

# With GPU acceleration
./llama-server \
    -m models/llama-2-70b.Q4_K_M.gguf \
    -ngl 40  # Offload 40 layers to GPU
```

--------------------------------

### Build for Production and Preview

Source: https://github.com/moltis-org/moltis/blob/main/docs/README.md

Navigate to the docs directory, build the project for production using npm run build, and preview the production build with npm run preview.

```bash
cd docs
npm run build
npm run preview
```

--------------------------------

### Install Moltis Service

Source: https://github.com/moltis-org/moltis/blob/main/docs/src/service.md

Installs and starts the Moltis service. This command creates the necessary service definition files and immediately launches the service.

```bash
moltis service install
```

--------------------------------

### Fetch Axolotl Examples

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/axolotl/references/other.md

Use this command to download example configurations and datasets for Axolotl. You can specify a custom destination path if needed.

```bash
axolotl fetch examples
```

```bash
axolotl fetch examples --dest path/to/folder
```

--------------------------------

### Start Moltis Gateway

Source: https://github.com/moltis-org/moltis/blob/main/website/llms.txt

Starts the Moltis gateway service, binding it to all network interfaces on a specified port. This is essential for agent communication and setup.

```bash
moltis gateway --bind 0.0.0.0 --port 13131
```

--------------------------------

### Skill Activation Example

Source: https://github.com/moltis-org/moltis/blob/main/docs/src/system-prompt.md

Illustrates how to activate a skill using the `read_skill` tool, with an example of loading a sidecar file.

```python
read_skill(name="<skill-name>")
read_skill(name="<skill-name>", file_path="references/api.md")
```

--------------------------------

### Install Apple Container

Source: https://github.com/moltis-org/moltis/blob/main/docs/src/sandbox.md

Download and install the Apple Container package, then perform the initial system setup to download a default Linux kernel.

```bash
# Download the installer package
gh release download --repo apple/container --pattern "container-installer-signed.pkg" --dir /tmp

# Install (requires admin)
sudo installer -pkg /tmp/container-installer-signed.pkg -target /

# First-time setup — downloads a default Linux kernel
container system start

```

--------------------------------

### Start Slidev Deck Locally

Source: https://github.com/moltis-org/moltis/blob/main/presentations/20260331/README.md

Starts the Slidev presentation deck locally and automatically opens it in the default browser. Ensure the Slidev CLI is installed.

```bash
slidev presentations/20260331/slides.md --open
```

--------------------------------

### Install Outlines and Backends

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/inference/outlines/SKILL.md

Install the base Outlines library and optional backends for specific model types like Hugging Face Transformers, llama.cpp, or vLLM.

```bash
pip install outlines

pip install outlines transformers  # Hugging Face models
pip install outlines llama-cpp-python  # llama.cpp
pip install outlines vllm  # vLLM for high-throughput
```

--------------------------------

### Install ascii-image-converter

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/creative/ascii-art/SKILL.md

Install the 'ascii-image-converter' tool using snap or go.

```bash
sudo snap install ascii-image-converter
# OR: go install github.com/TheZoraiz/ascii-image-converter@latest
```

--------------------------------

### Install and Enable Moltis systemd Service

Source: https://github.com/moltis-org/moltis/blob/main/docs/src/deploy-vps.md

Installs the Moltis systemd service file, reloads the systemd daemon, and enables and starts the Moltis service.

```bash
sudo curl -fsSL https://raw.githubusercontent.com/moltis-org/moltis/main/deploy/moltis.service -o /etc/systemd/system/moltis.service
sudo systemctl daemon-reload
sudo systemctl enable --now moltis
```

--------------------------------

### Run Qwen3-235B-A22B-Instruct using llama.cpp CLI

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/unsloth/references/llms-full.md

Executes the Qwen3-235B-A22B-Instruct model directly from the command line using llama.cpp. Adjust --threads, --ctx-size, and --n-gpu-layers for performance tuning based on your hardware.

```bash
./llama.cpp/llama-cli \
    -hf unsloth/Qwen3-235B-A22B-Instruct-2507-GGUF \
    --threads -1 \
    --ctx-size 16384 \
    --n-gpu-layers 99 \
    -ot ".ffn_.*_exps.=CPU" \
    --temp 0.7 \
    --min-p 0.0 \
    --top-p 0.8 \
    --top-k 20
```

--------------------------------

### PDF Processing with pdfplumber Example

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/software-development/skill-creator/SKILL.md

Demonstrates how to extract text from PDF files using the pdfplumber library. This is a quick start example for PDF processing.

```markdown
# PDF Processing

## Quick start

Extract text with pdfplumber:
[code example]
```

--------------------------------

### Setup Magistral Fine-tuning with Unsloth (Python)

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/unsloth/references/llms-full.md

Install Unsloth and load a pre-trained Magistral model for fine-tuning. This setup is optimized for memory efficiency and speed.

```python
!pip install --upgrade unsloth
from unsloth import FastLanguageModel
import torch
model, tokenizer = FastLanguageModel.from_pretrained(
    model_name = "unsloth/Magistral-Small-2509-unsloth-bnb-4bit",
    max_seq_length = 2048,   # Context length - can be longer, but uses more memory
    load_in_4bit = True,     # 4bit uses much less memory
    load_in_8bit = False,    # A bit more accurate, uses 2x memory
    full_finetuning = False, # We have full finetuning now!
    device_map = "balanced", # Uses 2x Telsa T4s
    # token = "hf_...",      # use one if using gated models
)
```

--------------------------------

### Configuration File Example

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/research/summarize/SKILL.md

Shows the structure of the optional configuration file for setting default models or other parameters. This file should be located at `~/.summarize/config.json`.

```json
{
  "model": "openai/gpt-5.2"
}
```

--------------------------------

### Attention Mechanism Visualization (Example 3)

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/unsloth/references/llms-txt.md

This example visualizes keys and queries in an attention mechanism, noting that q4 has a q_idx of 0 in the current setup.

```unknown
k0 k1 k2 k3 k4
q0
q1
q2
q3
q4   X   (note that q4 has q_idx=0 as this is the first query in current setup)
```

--------------------------------

### Explore Repository Command

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/research/research-paper-writing/SKILL.md

Use this bash command to begin exploring the project repository as part of the initial setup phase.

```bash
git ls-tree -r HEAD
```

--------------------------------

### Install Moltis Agent Server Script

Source: https://github.com/moltis-org/moltis/blob/main/website/index.en.html

This shell command downloads and executes the Moltis installation script. It is designed for quick setup on compatible systems.

```shell
$ curl -fsSL moltis.org/install.sh | sh
```

--------------------------------

### Dark Theme Iteration Guide

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/creative/popular-web-designs/templates/sanity.md

A step-by-step guide for building dark-themed interfaces, starting with base colors and progressing to typography, spacing, and technical details.

```text
1. **Start dark**: Begin with `#0b0b0b` background, `#ffffff` primary text, `#b9b9b9` secondary text
2. **Add structure**: Use `#212121` surfaces and `#353535` borders for containment -- no shadows
3. **Apply typography**: Inter (or Space Grotesk) with tight letter-spacing on headings, 1.50 line-height on body
4. **Color punctuation**: Add `#f36458` for CTAs and `#0052ef` for all hover/interactive states
5. **Refine spacing**: 8px base unit, 24-32px within sections, 64-120px between sections
6. **Technical details**: Add IBM Plex Mono uppercase labels for tags and metadata
7. **Polish**: Ensure all interactive elements hover to `#0052ef`, all buttons are pills or subtle 5px radius, borders are hairline (1px)
```

--------------------------------

### Basic Music Generation Example

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/media/heartmula/SKILL.md

Run the music generation script with specified model paths, lyrics, tags, and save path. Supports lazy loading for reduced VRAM usage.

```bash
cd heartlib
. .venv/bin/activate
python ./examples/run_music_generation.py \
  --model_path=./ckpt \
  --version="3B" \
  --lyrics="./assets/lyrics.txt" \
  --tags="./assets/tags.txt" \
  --save_path="./assets/output.mp3" \
  --lazy_load true
```

--------------------------------

### Complete Git Workflow Example

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/github/github-pr-workflow/SKILL.md

A comprehensive example demonstrating the typical steps for a Git workflow, from checking out main to creating a PR and merging.

```bash
# 1. Start from clean main
git checkout main && git pull origin main

# 2. Branch
git checkout -b fix/login-redirect-bug

# 3. (Agent makes code changes with file tools)

# 4. Commit
git add src/auth/login.py tests/test_login.py
git commit -m "fix: correct redirect URL after login

Preserves the ?next= parameter instead of always redirecting to /dashboard."

# 5. Push
git push -u origin HEAD

# 6. Create PR (picks gh or curl based on what's available)
# ... (see Section 3)

# 7. Monitor CI (see Section 4)

# 8. Merge when green (see Section 6)
```

--------------------------------

### Basic OpenAI Model Setup

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/inference/outlines/references/backends.md

Integrate with OpenAI models using their API key. This example shows basic setup and usage with an Outlines JSON generator.

```python
import outlines

# Basic OpenAI support
model = outlines.models.openai("gpt-4o-mini", api_key="your-api-key")

# Use with generator
generator = outlines.generate.json(model, YourModel)
result = generator("Your prompt")
```

--------------------------------

### setup_trainer

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/axolotl/references/api.md

Helper method for instantiating and building a trainer.

```APIDOC
## utils.trainer.setup_trainer

### Description
Helper method for instantiating and building a (causal or RLHF) trainer.

### Parameters
- **cfg**: Configuration object.
- **train_dataset**: The training dataset.
- **eval_dataset**: The evaluation dataset.
- **model**: The model to be trained.
- **tokenizer**: The tokenizer for the model.
- **processor**: The data processor.
- **total_num_steps**: The total number of training steps.
- **model_ref** (optional): Reference to another model. Defaults to None.
- **peft_config** (optional): PEFT configuration. Defaults to None.

### Example
```python
utils.trainer.setup_trainer(
    cfg,
    train_dataset,
    eval_dataset,
    model,
    tokenizer,
    processor,
    total_num_steps,
    model_ref=None,
    peft_config=None,
)
```
```

--------------------------------

### Skill Organization by Variant Example

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/software-development/skill-creator/SKILL.md

Shows how to organize a skill that supports multiple frameworks or variants by creating separate reference files for each. This allows for loading specific provider details as needed.

```directory structure
cloud-deploy/
├── SKILL.md (workflow + provider selection)
└── references/
    ├── aws.md (AWS deployment patterns)
    ├── gcp.md (GCP deployment patterns)
    └── azure.md (Azure deployment patterns)
```

--------------------------------

### Prerequisites: Install and Configure signal-cli

Source: https://github.com/moltis-org/moltis/blob/main/docs/src/signal.md

Before using Signal with Moltis, ensure signal-cli is installed and configured. This involves registering a number, verifying it, and starting the daemon in HTTP mode.

```bash
signal-cli -u +15551234567 register
signal-cli -u +15551234567 verify 123456
signal-cli --account +15551234567 daemon --http 127.0.0.1:8080
```

--------------------------------

### Release Prep Script Usage

Source: https://github.com/moltis-org/moltis/blob/main/plans/2026-02-17-calver-yyyy-m-ddnn-migration-plan.md

Demonstrates the command-line usage for the release preparation script, including version and release date arguments.

```bash
./scripts/prepare-release.sh 2026.2.1700 2026-02-17
```

--------------------------------

### Install Unsloth and vLLM for Fast Inference

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/unsloth/references/llms-full.md

Install the necessary libraries and initialize Unsloth with fast inference enabled. This setup allows for significantly improved throughput during finetuning and inference.

```python
pip install unsloth vllm
from unsloth import FastLanguageModel
model, tokenizer = FastLanguageModel.from_pretrained(
    model_name = "unsloth/Llama-3.2-3B-Instruct",
    fast_inference = True,
)
model.fast_generate(["Hello!"])
```

--------------------------------

### DSPy Evaluation Setup

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/research/dspy/SKILL.md

Demonstrates how to set up and use the DSPy evaluator with a development dataset and a custom metric. Shows how to evaluate a system and compare performance improvements.

```python
from dspy.evaluate import Evaluate

# Create evaluator
evaluator = Evaluate(
    devset=testset,
    metric=exact_match,
    num_threads=4,
    display_progress=True
)

# Evaluate model
score = evaluator(qa_system)
print(f"Accuracy: {score}")

# Compare optimized vs unoptimized
score_before = evaluator(qa)
score_after = evaluator(optimized_qa)
print(f"Improvement: {score_after - score_before:.2%}")
```

--------------------------------

### Install Core ML Dependencies

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/unsloth/references/llms-full.md

Install essential machine learning libraries including PyTorch, torchvision, torchaudio, and Triton. Ensure these versions are compatible with your CUDA setup and Unsloth requirements.

```bash
pip install torch torchvision torchaudio triton
```

--------------------------------

### Install Docker

Source: https://github.com/moltis-org/moltis/blob/main/docs/src/deploy-vps.md

Installs Docker on the server. After running this, log out and back in for group membership to take effect.

```bash
curl -fsSL https://get.docker.com | sh
sudo usermod -aG docker $USER
# Log out and back in for group membership to take effect
```

--------------------------------

### Axolotl Development Environment Setup

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/axolotl/references/other.md

Install necessary packages for Axolotl development, including packaging and optional dependencies like flash-attn and deepspeed. Use --no-build-isolation for cleaner installations.

```bash
pip3 install packaging
pip3 install --no-build-isolation -e '.[flash-attn,deepspeed]'
```

--------------------------------

### gog CLI Setup Commands

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/productivity/gog/SKILL.md

Commands to set up authentication and add services for the gog CLI. Ensure you have your client secret JSON file ready.

```bash
gog auth credentials /path/to/client_secret.json
```

```bash
gog auth add you@gmail.com --services gmail,calendar,drive,contacts,docs,sheets
```

```bash
gog auth list
```

--------------------------------

### Browser Automation Workflow Example

Source: https://github.com/moltis-org/moltis/blob/main/docs/src/browser-automation.md

A comprehensive example illustrating a typical browser automation workflow, including navigating to a page, taking a snapshot to get element references, filling a form, clicking a button, and taking a screenshot.

```json
// 1. Navigate to a page
{
  "action": "navigate",
  "url": "https://example.com/login"
}
// Returns: { "session_id": "browser-abc123", "url": "https://..." }

// 2. Get interactive elements (session_id optional - auto-tracked)
{
  "action": "snapshot"
}
// Returns element refs like:
// { "elements": [
//   { "ref_": 1, "tag": "input", "role": "textbox", "placeholder": "Email" },
//   { "ref_": 2, "tag": "input", "role": "textbox", "placeholder": "Password" },
//   { "ref_": 3, "tag": "button", "role": "button", "text": "Sign In" }
// ]}

// 3. Fill in the form
{
  "action": "type",
  "ref_": 1,
  "text": "user@example.com"
}

{
  "action": "type",
  "ref_": 2,
  "text": "password123"
}

// 4. Click the submit button
{
  "action": "click",
  "ref_": 3
}

// 5. Take a screenshot of the result
{
  "action": "screenshot"
}
// Returns: { "screenshot": "data:image/png;base64,வைக்" }
```

--------------------------------

### Create and Activate Virtual Environment with Dependencies

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/media/heartmula/SKILL.md

Set up a Python 3.10 virtual environment using uv, activate it, and install HeartLib and its dependencies.

```bash
uv venv --python 3.10 .venv
. .venv/bin/activate
uv pip install -e .
```

--------------------------------

### Install QMD CLI

Source: https://github.com/moltis-org/moltis/blob/main/docs/src/memory.md

Install the QMD CLI using npm or bun. Ensure the binary is on your PATH.

```bash
npm install -g @tobi-lu/qmd
```

```bash
bun install -g @tobi-lu/qmd
```

--------------------------------

### End-to-End Testing Setup

Source: https://github.com/moltis-org/moltis/blob/main/docs/src/local-validation.md

The `local/e2e` check automatically runs `npm ci` if `crates/web/ui/node_modules` is missing, followed by `npm run e2e:install` and `npm run e2e`. This ensures the UI dependencies are installed and the end-to-end tests are executed.

```bash
npm ci
npm run e2e:install
npm run e2e
```

--------------------------------

### Install Unsloth, Unsloth Zoo, and bitsandbytes with uv

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/unsloth/references/llms-txt.md

Install Unsloth, Unsloth Zoo, and bitsandbytes packages using uv.

```bash
uv pip install unsloth unsloth_zoo bitsandbytes
```

--------------------------------

### Train Diffusion Language Model (Pretrain)

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/axolotl/references/other.md

Example command to start pre-training a diffusion language model using Axolotl.

```bash
axolotl train examples/llama-3/diffusion-3.2-1b-pretrain.yaml
```

--------------------------------

### Camera-Ready Version Setup

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/research/research-paper-writing/templates/aaai2026/README.md

Configuration for the camera-ready version by commenting out or deleting the \def\aaaianonymous macro. Add complete author and affiliation information.

```latex
% Commented out \def\aaaianonymous{true} (camera-ready)
```

--------------------------------

### Train Diffusion Language Model (SFT)

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/axolotl/references/other.md

Example command to start supervised fine-tuning (SFT) of a diffusion language model using Axolotl.

```bash
axolotl train examples/llama-3/diffusion-3.2-1b-sft.yaml
```

--------------------------------

### Review Strategy Commands

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/github/github-code-review/SKILL.md

Commands to get a high-level overview of changes, including file statistics and commit history, to guide a detailed review.

```bash
git diff main...HEAD --stat
git log main..HEAD --oneline
```

--------------------------------

### Fetch Resources with CLI

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/axolotl/references/api.md

Retrieve example configurations or other resources using the Axolotl CLI's fetch command.

```python
cli.main.fetch(directory, dest)
```

--------------------------------

### Pre-training Dataset Configuration Example

Source: https://github.com/moltis-org/moltis/blob/main/crates/skills/src/assets/mlops/training/axolotl/references/dataset-formats.md

Example configuration for pre-training, including dataset source and optional parameters like `group_by_length`.

```yaml
train_dataset:
  hf_dataset:
    repo: "TheBloke/vicuna-7B-v1.5-GPTQ"
    split: "train"
  group_by_length: false

```