- Home
- Skills
- Outfitter Dev
- Agents
- Claude Commands
claude-commands_skill
- TypeScript
25
GitHub Stars
2
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill outfitter-dev/agents --skill claude-commands- EXAMPLES.md20.7 KB
- SKILL.md10.5 KB
Overview
This skill provides a concise guide and conventions for authoring custom slash commands for Claude Code. It explains file layout, frontmatter schema, argument patterns, bash execution, file references, and validation checklists. Use it to create reproducible, safe, and discoverable project or personal commands. The guidance focuses on practical rules and examples so commands work reliably in team and local contexts.
How this skill works
Commands are single Markdown files placed in a commands directory and invoked explicitly with a slash (e.g., /review or /fix-issue 123). Each command uses YAML frontmatter to declare description, allowed tools, model overrides, and argument hints; content can reference positional arguments ($1, $2), $ARGUMENTS, file inclusions (@path), and preprocessed bash output (!command). After authoring, commands are discovered via /help and validated against naming, frontmatter, and content quality checks.
When to use it
- When you need a reusable, user-invoked prompt or workflow (slash command).
- When automating repo-specific tasks like reviews, commits, or deployments.
- When you want to expose controlled tool access (Bash, Read, Grep) for a command.
- When you need argument parsing or file inclusion in a prompt.
- When bundling shareable commands for a project, plugin, or personal use.
Best practices
- Use concise, action-oriented descriptions under 80 characters for /help discovery.
- Name files in kebab-case without spaces (e.g., fix-issue.md).
- Declare allowed-tools explicitly to limit runtime capabilities and increase safety.
- Prefer positional arguments ($1, $2) or $ARGUMENTS and document argument-hint in frontmatter.
- Test bash snippets in a terminal before embedding; prefix backticks with ! for preprocessing.
- Validate commands with the provided checklist: frontmatter, naming, content, and tool syntax.
Example use cases
- Create /review to check code quality, potential bugs, and security issues for a PR.
- Create /fix-issue <number> to fetch an issue reference, implement a fix, and produce a commit message.
- Create /commit to generate a git commit message from staged changes using Bash context.
- Create deployment commands that run pre-flight checks and optionally skip tests via flags.
- Bundle commands into a plugin folder to distribute common workflows to other projects.
FAQ
Project commands belong in .claude/commands/ (shared via git). Personal commands go in ~/.claude/commands/ and are private to your account.
How do I pass arguments with spaces?
Quote arguments when invoking (e.g., /cmd "arg with spaces"). Use $ARGUMENTS to access the whole argument string.
How do I include shell output in the command context?
Prefix backticks with an exclamation mark like !git status --short to run the command and embed its output.