Guide · Agent skill

How to use the Create Forge Commands agent skill

Updated 2026-11-02 · written with AI from the project's own documentation and checked against our catalogue data

The create-command skill is a set of instructions that teaches an AI agent how to write custom commands for the code-forge coding tool. It is for developers who want repeatable workflows, such as a pre-commit check or a PR description generator, stored inside their project. After reading this guide you will know how to install the skill, ask for a command in plain language, and check the result before running it.

Create Forge Commandstailcallhq/forgecode · Apache-2.0 Download the skill

Quick start

1Download the skill ZIP. Download the ZIP from this page and unzip it so that SKILL.md is available.
2Place the skill folder. Move the folder into ~/.claude/skills/ for all projects, or into your project's .claude/skills/ folder for one project. In the Claude apps, upload it under Settings → Capabilities → Skills instead.
3Make sure code-forge is installed. The commands only run inside the code-forge application, so install it before you start.
4Ask for a command in plain language. Describe what the command should do and which checks it should run.
5Check the file location. Confirm the agent wrote the file to <cwd>/.forge/commands/ with a name like command-name.md.
6Review every tag. Read each <lint>, <test> and <shell> line and confirm the tools it calls are ones you want to run.
7Run and refine. Invoke the command in code-forge, see what it does, and ask the agent to adjust steps or tags.

What an agent skill is and what this one does

An agent skill is a folder containing a SKILL.md file. The file gives an AI agent background instructions it can use when a task matches the skill's description. This skill covers one job: creating and editing code-forge command files.

A code-forge command is a markdown file with a name, a description and a list of steps. The skill tells the agent where the file must go, how to format it, how to choose a name, and how to use special tags such as lint, test and shell.

Saves each command as {command-name}.md in the project's .forge/commands directory
Writes YAML frontmatter with name and description fields
Adds steps as a markdown list, optionally with <lint>, <test> or <shell> tags

Install the create-command skill

Download the ZIP from this page and unzip it. The skill is contained in SKILL.md. Then place the folder in one of two locations: ~/.claude/skills/ to make it available across your projects, or a project's .claude/skills/ folder to make it available only in that project.

If you use the Claude apps, you can instead upload the skill under Settings → Capabilities → Skills.

Personal install: ~/.claude/skills/
Project install: the .claude/skills/ folder inside your project
Claude apps: Settings → Capabilities → Skills, then upload

Trigger the skill with natural requests

You do not need a special command to use the skill. Ask the agent in plain language, and it will load the skill when your request matches its description. The description covers adding new commands, modifying existing commands, and understanding the command file structure.

Name the outcome you want and, if you have one, the checks it should run. The agent will then decide the filename and frontmatter.

"Create a code-forge command called pr-description that updates the PR title and body with the GH CLI."
"Add a command that runs lint and tests before I commit."
"Explain the file structure for forge commands in this repo."
"Edit the existing check command so it also runs cargo test."

Understand the command file format

Every command file needs a frontmatter block and a body. The name uses hyphens for multi-word names, and the description says what the command does. The body lists the steps. Tags go on their own line directly after the step they belong to.

The skill's example command, check, runs lint and test and then asks the agent to fix any issues found.

Location: <cwd>/.forge/commands/{command-name}.md, where <cwd> is your project's working directory
Naming: lowercase, hyphens between words, verb-based (for example check, fixme, pr-description)
Tags: <lint>, <test> and <shell>; the skill says to use only these three

---
name: check
description: Checks if the code is ready to be committed
---

- Run the `lint` and `test` commands and verify if everything is fine.
  <lint>cargo +nightly fmt --all; cargo +nightly clippy --fix --allow-staged --allow-dirty --workspace</lint>
  <test>cargo insta test --accept --unreferenced=delete</test>
- Fix every issue found in the process

Review tags before running any command

Tags are not just text. According to the skill, <lint>, <test> and <shell> cause the listed shell commands to be run on your machine. The commands in the examples come from the skill and from the Rust toolchain and git; they are not safe defaults for every repository.

Before you run a command in a repository you did not write, read every tag and check the shell commands it contains. The skill declares that it runs scripts, and it requires shell tools such as cargo, git or GitHub CLI, depending on what the command runs.

Read every <lint>, <test> and <shell> line before running the command
Confirm the tools (cargo, git, GitHub CLI) are installed and that you expect them to run
Be cautious with destructive commands, such as the rm -rf example in the skill's cleanup patterns
The skill does not use the network itself, but the commands it writes may call tools that do

Templates and patterns the skill provides

The skill includes templates for a simple command, an automated workflow, a multi-step workflow and a git workflow. It also lists common patterns such as commit checks, documentation updates and cleanup. Use them as starting points and edit the steps to match your project.

Simple command: one instruction in the body
Automated workflow: lint and test tags followed by follow-up actions
Git workflow: stage, run checks, commit and push with shell tags
Cleanup: removes build artifacts with a shell tag

What people use it for

Create a check command that runs lint and test before a commit.
Write a pr-description command that uses the GH CLI to update a pull request title and body.
Add a cleanup command that removes build artifacts with a shell tag.
Edit an existing command to add a step or change its description.
Share a standard workflow command through your repository's .forge/commands folder.

Frequently asked questions

What is the create-command skill?

It is an agent skill that teaches an AI agent how to write and edit custom command files for the code-forge application. Commands are markdown files with YAML frontmatter, saved in the project's .forge/commands folder.

Do I need code-forge to use this skill?

Yes. The skill's commands only work in the code-forge application, so the skill is not useful without it.

Is the create-command skill free?

The source is licensed under Apache-2.0, which is listed for this skill. The source does not state any price, so check the repository for licence terms.

Where do I put the skill so my agent can use it?

Place the skill folder in ~/.claude/skills/ for all projects, or in a project's .claude/skills/ folder. In the Claude apps, upload it under Settings → Capabilities → Skills.

Why isn't forge finding my command?

Forge only loads commands from <cwd>/.forge/commands. Check that the file is in that folder, uses the {command-name}.md format, and has valid frontmatter with name and description.

Are there safer alternatives to running tags blindly?

Yes. Review each tag before running a command, and only use the lint, test and shell tags for commands you understand. The skill does not offer a sandbox, so keep commands to a reviewed set in repositories you trust.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Choose this skill if you use code-forge and want an agent to write or explain its command files for your project. Skip it if you do not use code-forge, or if you plan to run commands from a repository whose tags you have not reviewed.

Download the skill