### Install and Setup Modal CLI

Source: https://github.com/companion-inc/feynman/blob/main/skills/modal-compute/SKILL.md

Install the Modal CLI and complete the setup process before running any commands.

```bash
pip install modal
modal setup
```

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

### Run Feynman Setup Wizard

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/setup.md

Execute the command to start the interactive setup wizard for configuring Feynman.

```bash
feynman setup
```

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

### Install Optional Packages

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/setup.md

Skip package installation during setup and install them later using this command.

```bash
feynman packages install <preset>
```

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

### Install Preview Dependencies

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/tools/preview.md

Install the necessary dependencies for the preview tool, including pandoc, using the feynman setup command.

```bash
feynman setup preview
```

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

### Install Dependencies and Use Node.js Version

Source: https://github.com/companion-inc/feynman/blob/main/CONTRIBUTING.md

Installs the correct Node.js version and then installs project dependencies from the repository root. Ensure you have nvm installed and configured.

```bash
nvm use || nvm install
npm install
```

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

### Clone and Install Feynman

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Clone the Feynman repository, navigate to the directory, install Node.js version, and install dependencies.

```bash
git clone https://github.com/companion-inc/feynman.git
cd feynman
nvm use || nvm install
npm install
```

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

### Install Feynman Skills for Repo (Windows)

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Installs Feynman skills for a repository-local Claude/agent install on Windows.

```powershell
& ([scriptblock]::Create((irm https://feynman.is/install-skills.ps1))) -Scope Repo
```

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

### Install Feynman Skills for OpenCode (Windows)

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Installs Feynman skills for an OpenCode project-local install on Windows.

```powershell
& ([scriptblock]::Create((irm https://feynman.is/install-skills.ps1))) -Scope OpenCode
```

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

### Show Current Setup Summary

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/reference/cli-commands.md

Display a summary of the current Feynman setup, including model, authentication, and packages.

```bash
feynman status
```

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

### Alias for `feynman setup preview`

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/reference/cli-commands.md

An alias flag that executes the `feynman setup preview` command.

```bash
--setup-preview
```

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

### Install Feynman Skills for OpenCode (macOS/Linux)

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Installs Feynman skills for an OpenCode project-local install on macOS or Linux.

```bash
curl -fsSL https://feynman.is/install-skills | bash -s -- --opencode
```

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

### Install and Configure runpodctl

Source: https://github.com/companion-inc/feynman/blob/main/skills/runpod-compute/SKILL.md

Install the runpodctl CLI using Homebrew and configure it with your API key. Ensure you replace YOUR_KEY with your actual RunPod API key.

```bash
brew install runpod/runpodctl/runpodctl
runpodctl config --apiKey=YOUR_KEY
```

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

### Install Feynman Skills for Repo (macOS/Linux)

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Installs Feynman skills for a repository-local Claude/agent install on macOS or Linux.

```bash
curl -fsSL https://feynman.is/install-skills | bash -s -- --repo
```

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

### Install Session Search Package

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/tools/session-search.md

Install the session search tool as an optional package using the feynman CLI.

```bash
feynman packages install session-search
```

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

### List Available Packages

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/reference/package-stack.md

Lists all available packages and their current installation status within the Feynman environment.

```bash
feynman packages list
```

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

### Enter API Key

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/setup.md

When prompted by the setup wizard for providers requiring an API key, paste your key here. Keys are encrypted.

```bash
? Enter your API key: sk-ant-...
```

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

### Install Feynman CLI with Specific Version

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Installs a specific version of the Feynman CLI by passing the version number to the installer.

```bash
curl -fsSL https://feynman.is/install | bash -s -- 0.2.35
```

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

### Launch Feynman REPL

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/quickstart.md

Start an interactive conversational REPL session to ask research questions and interact with agents.

```bash
feynman
```

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

### Find ML Training Recipe

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/quickstart.md

Use the 'recipe' command to find practical starting points for applied ML tasks, such as fine-tuning a model. The output is saved to a markdown file.

```bash
feynman recipe "fine-tune a small model for math reasoning"
```

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

### Install Optional Package Preset

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/reference/package-stack.md

Installs a specific optional package preset, such as `generative-ui`, to extend Feynman's capabilities.

```bash
feynman packages install generative-ui
```

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

### Start Deep Research Session

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/quickstart.md

Initiate a deep research workflow within the REPL to get a structured report on a topic, complete with citations and findings.

```bash
feynman
> /deepresearch What are the current approaches to mechanistic interpretability in LLMs?
```

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

### Install Feynman CLI (Windows)

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Installs the latest tagged release of the Feynman CLI using PowerShell.

```powershell
irm https://feynman.is/install.ps1 | iex
```

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

### Install Feynman (Latest Release)

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/installation.md

Use this command to install the latest tagged release of Feynman on Unix-like systems.

```bash
curl -fsSL https://feynman.is/install | bash -s --
```

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

### Start Chat with Initial Prompt

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/reference/cli-commands.md

Initiate a chat session explicitly, optionally providing an initial prompt.

```bash
feynman chat [prompt]
```

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

### Install Feynman CLI (macOS/Linux)

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Installs the latest tagged release of the Feynman CLI using curl.

```bash
curl -fsSL https://feynman.is/install | bash
```

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

### Feynman Command: Get ML Recipe

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Finds ranked, implementable ML training recipes from various sources.

```bash
$ feynman recipe "fine-tune a small model for math reasoning"
→ Finds ranked, implementable ML training recipes from papers, datasets, docs, and code
```

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

### Install Feynman Skills for Codex (Windows)

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Installs Feynman skills specifically for Codex on Windows.

```powershell
& ([scriptblock]::Create((irm https://feynman.is/install-skills.ps1))) -Scope Codex
```

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

### Verify Feynman Installation

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/installation.md

Confirm that Feynman has been installed correctly and is accessible by checking its version number.

```bash
feynman --version
```

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

### Install Feynman Skills Only (Windows)

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Installs only the research skills for Feynman on Windows.

```powershell
irm https://feynman.is/install-skills.ps1 | iex
```

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

### Install Feynman Skills Only (macOS/Linux)

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Installs only the research skills for Feynman on macOS or Linux.

```bash
curl -fsSL https://feynman.is/install-skills | bash
```

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

### Verify Feynman Configuration

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/configuration.md

Run this command to check the validity of your Feynman setup, including authentication status and dependency checks.

```bash
feynman doctor
```

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

### Run Literature Review from REPL

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/workflows/literature-review.md

Initiate a literature review by providing the topic directly in the REPL. This is a quick way to start the process.

```bash
/lit Scaling laws for language model performance
```

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

### Install Feynman Skills for Codex (macOS/Linux)

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Installs Feynman skills specifically for Codex on macOS or Linux.

```bash
curl -fsSL https://feynman.is/install-skills | bash -s -- --codex
```

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

### Start a New Session via CLI

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/configuration.md

Use this flag to initiate a completely fresh conversation session, discarding any previous state or history.

```bash
feynman --new-session
```

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

### Start Autoresearch from REPL

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/workflows/autoresearch.md

Initiate an autoresearch task from the REPL by providing a clear objective. This command starts a long-lived background process for iterative optimization.

```bash
/autoresearch Optimize prompt engineering strategies for math reasoning on GSM8K
```

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

### Install Feynman via npm

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/installation.md

Install Feynman as an npm package if you prefer using an existing Node.js environment. Ensure your Node.js version is compatible.

```bash
npm install -g @companion-ai/feynman
```

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

### Start New Persisted Session

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/reference/cli-commands.md

Initiate a new session that will be persisted, allowing state to be saved and resumed.

```bash
--new-session
```

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

### Start Autoresearch Loop

Source: https://github.com/companion-inc/feynman/blob/main/prompts/autoresearch.md

Initiates the autonomous research and optimization loop. Use this command to begin the process of trying ideas, measuring results, and iterating on improvements.

```bash
autoresearch <idea>
```

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

### Fetch and Query Academic Papers with Alpha CLI

Source: https://github.com/companion-inc/feynman/blob/main/skills/alpha-research/SKILL.md

Use `alpha get` to fetch paper content. Use `alpha ask` to ask questions about a specific paper identified by its arXiv ID.

```bash
alpha get 2106.09685
```

```bash
alpha ask 2106.09685 "What optimizer did they use?"
```

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

### Start Autoresearch from CLI

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/workflows/autoresearch.md

Execute an autoresearch task using the Feynman CLI. This is useful for scripting or running autoresearch from a terminal.

```bash
feynman autoresearch "Optimize prompt engineering strategies for math reasoning on GSM8K"
```

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

### RLM-lite Windowed Read Example

Source: https://github.com/companion-inc/feynman/blob/main/prompts/summarize.md

Demonstrates how to extract bounded windows of text from a file using bash for the RLM-lite strategy. This method keeps the document on disk and avoids injecting raw content into the context.

```python
# WHY f.seek/f.read: the read tool uses line offsets, not char offsets.

```

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

### Manage RunPod Pods

Source: https://github.com/companion-inc/feynman/blob/main/skills/runpod-compute/SKILL.md

Common commands to manage your RunPod pods, including listing, stopping, starting, and removing them. Use the pod ID obtained from 'runpodctl get pod'.

```bash
runpodctl get pod
runpodctl stop pod <id>
runpodctl start pod <id>
runpodctl remove pod <id>
```

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

### Create and Manage a Persistent Docker Container

Source: https://github.com/companion-inc/feynman/blob/main/skills/docker/SKILL.md

Create a named Docker container for iterative experiments, allowing installed packages to persist across runs. Start, execute commands within, and stop/remove the container.

```bash
docker create --name <name> -v "$(pwd)":/workspace -w /workspace python:3.11 tail -f /dev/null
docker start <name>
docker exec <name> bash -c "pip install -r requirements.txt"
docker exec <name> bash -c "python train.py"
docker stop <name> && docker rm <name>
```

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

### Show CLI Help

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/reference/cli-commands.md

Display the help message for the Feynman CLI, listing available commands and options.

```bash
feynman help
```

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

### Install Specific Feynman Version (Windows)

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/installation.md

On Windows, use PowerShell to install a specific version of Feynman by passing the version number to the install script.

```powershell
& ([scriptblock]::Create((irm https://feynman.is/install.ps1))) -Version 0.2.31
```

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

### Run Recipe Workflow from REPL

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/workflows/recipe.md

Initiate the recipe workflow from the REPL to find implementable ML training recipes for a specified goal.

```bash
/recipe "fine-tune a small model for math reasoning"
```

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

### Initialize Experiment Session

Source: https://github.com/companion-inc/feynman/blob/main/prompts/autoresearch.md

Configures a new experiment session, including setting the session name, optimization metric, its unit, and the direction of improvement (lower or higher is better).

```bash
init_experiment
```

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

### Replicate from CLI

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/workflows/replication.md

Execute the replication workflow from the command line interface by providing the paper or claim as an argument.

```bash
feynman replicate "paper or claim"
```

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

### Uninstall Feynman installed via npm

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/installation.md

Use this npm command to uninstall Feynman if it was installed as an npm package.

```bash
npm uninstall -g @companion-ai/feynman
```

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

### Set Web Search Provider and API Key

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/tools/web-search.md

This command-line interface command can be used to write the web search configuration file. It allows setting the provider and optionally an API key.

```bash
feynman search set <provider> [api-key]
```

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

### List Configured Models via CLI

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/configuration.md

This command displays all AI models that are currently authenticated and configured for use with Feynman.

```bash
feynman model list
```

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

### Inappropriate Prompt Examples

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/workflows/biomedical-literature-review.md

Examples of prompts that should be avoided when using Feynman for biomedical topics, as they solicit direct medical advice or diagnosis.

```text
What should this patient do?
Which treatment should I choose for this case?
Diagnose this patient.
```

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

### Create a RunPod GPU Pod

Source: https://github.com/companion-inc/feynman/blob/main/skills/runpod-compute/SKILL.md

Create a new GPU pod with a specified GPU type, Docker image, and a name for your experiment. This command provisions the compute resources.

```bash
runpodctl create pod --gpuType "NVIDIA A100 80GB PCIe" --imageName "runpod/pytorch:2.4.0-py3.11-cuda12.4.1-devel-ubuntu22.04" --name experiment
```

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

### Configure Local Ollama Server

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/setup.md

Settings for connecting to a local Ollama server. Ensure Ollama is running and a model is loaded.

```text
API mode: openai-completions
Base URL: http://localhost:11434/v1
Authorization header: No
Model ids: llama3.1:8b
API key: local
```

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

### Update All Installed Packages

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/reference/package-stack.md

Updates all installed Pi packages within Feynman's environment to their latest available versions. This command is safe and preserves existing configurations.

```bash
feynman update
```

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

### Research Question Framing Examples

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/workflows/biomedical-literature-review.md

Examples of how to phrase research tasks for Feynman when dealing with sensitive topics like diagnosis or treatment. Focus on literature synthesis rather than direct advice.

```text
Summarize the research literature on...
Compare the evidence for...
Identify limitations and gaps in studies about...
Draft a background section for a research protocol on...
```

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

### Preview Specific File

Source: https://github.com/companion-inc/feynman/blob/main/skills/preview/SKILL.md

Use the /preview --file command to render and open a specific artifact file in the browser.

```bash
/preview --file <path>
```

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

### Update Packages

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/reference/cli-commands.md

Update installed packages. Specify a package name to update a single package.

```bash
feynman update [package]
```

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

### Feynman Command: Replicate Experiment

Source: https://github.com/companion-inc/feynman/blob/main/README.md

Replicates experiments on local or cloud GPUs based on a given prompt.

```bash
$ feynman replicate "chain-of-thought improves math"
→ Replicates experiments on local or cloud GPUs
```

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

### Update Feynman via npm

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/installation.md

If you installed Feynman with npm, use this command to upgrade it to the latest version.

```bash
npm install -g @companion-ai/feynman@latest
```

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

### Set Up Watch from CLI

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/workflows/watch.md

Initiates a recurring research monitoring task for a specified topic using the command-line interface.

```bash
feynman watch "New developments in state space models for sequence modeling"
```

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

### Authenticate with AlphaXiv

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/tools/alphaxiv.md

Use this command to log in to AlphaXiv. Authentication tokens are stored locally.

```bash
feynman alpha login
```

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

### Compare topics using CLI

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/workflows/compare.md

Use the command-line interface to compare sources based on a topic or a list of specific source IDs.

```bash
feynman compare "topic or list of sources"
```

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

### Authenticate Model Provider

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/reference/cli-commands.md

Authenticate a model provider using OAuth or API-key setup. Specify the model ID if needed.

```bash
feynman model login [id]
```

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

### Run Commands Using a Project's Dockerfile

Source: https://github.com/companion-inc/feynman/blob/main/skills/docker/SKILL.md

Build a Docker image from a Dockerfile in the current directory and run a container, mounting a local 'results' directory to '/workspace/results'.

```bash
docker build -t feynman-experiment .
docker run --rm -v "$(pwd)/results":/workspace/results feynman-experiment
```

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

### List Available GPU Types

Source: https://github.com/companion-inc/feynman/blob/main/skills/runpod-compute/SKILL.md

View the available GPU types and their associated prices on RunPod. This helps in selecting the most suitable and cost-effective GPU for your workload.

```bash
runpodctl gpu list
```

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

### Login to AI Providers via CLI

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/configuration.md

Authenticate with different AI providers to make their models available for use in Feynman. This command initiates the respective authentication flows.

```bash
feynman model login anthropic
```

```bash
feynman model login google
```

```bash
feynman model login amazon-bedrock
```

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

### Preview Most Recent Artifact

Source: https://github.com/companion-inc/feynman/blob/main/skills/preview/SKILL.md

Use the /preview command to render and open the most recent artifact in the browser.

```bash
/preview
```

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

### AlphaXiv Login and Exit

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/reference/cli-commands.md

Sign in to AlphaXiv and immediately exit the CLI.

```bash
--alpha-login
```

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

### Clear Autoresearch State

Source: https://github.com/companion-inc/feynman/blob/main/prompts/autoresearch.md

Deletes all state and data associated with the current autoresearch loop, allowing for a completely fresh start.

```bash
autoresearch clear
```

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

### Update Specific Package

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/reference/package-stack.md

Updates a single, specified package (e.g., `pi-subagents`) to its latest version while leaving other installed packages unchanged.

```bash
feynman update pi-subagents
```

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

### Confirm and Set Local Model

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/getting-started/setup.md

After configuring a local provider, list available models and set your desired default.

```bash
feynman model list
feynman model set <provider>/<model-id>
```

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

### Run Peer Review from REPL

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/workflows/review.md

Initiate the peer review workflow from the REPL by providing an arXiv ID or a local file path.

```bash
/review arxiv:2401.12345
```

```bash
/review ~/papers/my-draft.pdf
```

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

### Replicate from REPL using Paper ID

Source: https://github.com/companion-inc/feynman/blob/main/website/src/content/docs/workflows/replication.md

Use this command in the REPL to start a replication workflow for a specific paper identified by its arXiv ID.

```bash
/replicate arxiv:2401.12345
```

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

### Force Browser Preview

Source: https://github.com/companion-inc/feynman/blob/main/skills/preview/SKILL.md

Use the /preview-browser command to force the artifact to be previewed in the browser.

```bash
/preview-browser
```

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

### Run GPU Workloads in a Docker Container

Source: https://github.com/companion-inc/feynman/blob/main/skills/docker/SKILL.md

Execute Python scripts on a GPU-enabled Docker image, installing dependencies first. Requires Docker to be configured with the NVIDIA Container Toolkit.

```bash
docker run --rm --gpus all -v "$(pwd)":/workspace -w /workspace pytorch/pytorch:latest bash -c "
  pip install -r requirements.txt &&
  python train.py
"
```