What does this skill do, and when should you use it?
This skill is for developers who create or maintain slash commands in Claude Code. It covers Markdown command files, YAML frontmatter, dynamic arguments, file references, inline Bash execution, and command namespaces. It also explains plugin commands, CLAUDE_PLUGIN_ROOT, and coordination with agents, skills, hooks, scripts, and templates. Because it depends on Claude Code’s slash-command behavior, it is best suited to Claude Code plugin development rather than general command scripting.
Explains how to organize Markdown commands in .claude/commands/, ~/.claude/commands/, or a plugin’s commands/ directory; configure description, allowed-tools, model, argument-hint, and disable-model-invocation frontmatter; use $ARGUMENTS and positional parameters such as $1; reference files with @; execute Bash inline with !...; and validate arguments, files, plugin resources, and failures.
- A Claude Code developer creating shared project commands such as /review or /deploy.
- A developer maintaining personal commands for workflows across multiple projects.
- A plugin author organizing bundled commands, namespaces, and portable plugin resources.
- An engineer building commands that accept arguments, read files, or run commands such as git, npm, or gh.
- A plugin author combining commands with agents, skills, scripts, and templates in multi-stage workflows.
How do you install this skill?
- Do not treat Bash(*), deployment, deletion, or external-service examples as safe defaults; narrow permissions and require explicit confirmation before side effects.
- Verify the documented frontmatter, $ARGUMENTS, @, !`, and AskUserQuestion syntax against the current Claude Code version before use; reproducible test evidence is not supplied.
- Resolve the 0.2.0 versus v0.1.0 version conflict and document dependencies, non-fit boundaries, sensitive-data handling, rollback, and mainland-China reachability.
- Shell / CLI
- Local filesystem
This skill belongs to the plugin-dev plugin in anthropics/claude-plugins-official. According to the repository README, run this in Claude Code: /plugin install plugin-dev@claude-plugins-official. The source does not document installation steps for other clients.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/anthropics/claude-plugins-official.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/plugins/plugin-dev/skills/command-development" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- create a slash command
- define command arguments.
- use command frontmatter,
- create command with file references,
After installation, use a trigger request such as “create a slash command” or “define command arguments.” Other documented triggers include “use command frontmatter,” “create command with file references,” and “interactive command.” The skill then provides guidance on command structure, arguments, tool permissions, file references, and Bash execution.
What are this skill's strengths and limitations?
- Covers topics from basic Markdown commands through plugin command organization.
- Includes examples for arguments, file references, Bash execution, validation, and error handling.
- Clearly distinguishes project, personal, and plugin command locations.
- Explains CLAUDE_PLUGIN_ROOT and integration with other plugin components.
- It is explicitly centered on Claude Code slash-command behavior and requires adaptation elsewhere.
- The source provides no test suite, platform compatibility matrix, or execution results.
- Bash examples may rely on external commands such as git, npm, gh, or node, without a consolidated dependency list.
- Some detailed specifications point to reference files whose contents are not included here.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Claude Code Command Builder this page ✓ Anthropic · Official | 46 · Use with care | ★ 38k | 11d ago | Apache-2.0 |
| Claude Code Command Development ✓ Anthropic · Official | 46 · Use with care | ★ 151k | 11d ago | — |
| Project Plugin Settings Pattern ✓ Anthropic · Official | 49 · Use with care | ★ 38k | 11d ago | Apache-2.0 |
| Claude Code Agent Development Guide ✓ Anthropic · Official | 54 · Use with care | ★ 38k | 11d ago | Apache-2.0 |
| Claude Code Plugin Skill Development ✓ Anthropic · Official | 51 · Use with care | ★ 38k | 11d ago | Apache-2.0 |
How did FollowSkills review this skill?
The documentation recommends allowed-tools restrictions, input validation, failure handling, and user confirmation, and warns against destructive operations. However, several examples use Bash(*), deployment scripts, rm, credentials, and external services without systematic data-flow disclosure, sensitive-data guidance, rollback boundaries, or confirmation requirements. No malware, credential theft, covert exfiltration, or other red-line risk is evident.
Core concepts, arguments, file references, validation, and error handling are covered with many reusable examples. Reliability is reduced because some examples use pseudocode-like conditionals, potentially ambiguous shell interpolation and execution semantics, and dependencies on scripts or reference files not supplied in this material. No verifiable execution or test evidence is present, so the static ceiling applies.
The frontmatter provides precise trigger phrases covering command creation, arguments, references, Bash, interaction, and best practices, and the documentation distinguishes project, personal, and plugin scope. Non-fit boundaries, runtime version requirements, and Chinese-language guidance are limited. Examples depend on Bash, npm, gh, kubectl, or remote services whose availability, including from mainland China, is not addressed.
The structure is reasonably progressive: SKILL.md, README, references, and examples provide layered guidance, troubleshooting, maintenance notes, version information, and update advice. However, frontmatter version 0.2.0 conflicts with the README's v0.1.0 history; the README's inventory does not fully match the supplied material; license and maintenance provenance are supplied mainly as metadata rather than clearly documented in the skill; and in-progress items lack a concrete update path.
The skill covers command structure, common frontmatter, dynamic arguments, file references, plugin-root usage, validation, and workflow patterns. The examples are directly reusable as templates and support the core task. It is not full-score evidence because examples were not executed, some syntax or tool behavior may require correction, and complex deployment patterns depend on external scripts, permissions, and environment-specific setup.
The supplied material includes auditable primary documentation, examples, checklists, and maintenance statements supporting the main claims. It lacks committed test suites, CI results, versioned behavior comparisons, or third-party execution evidence, and some referenced files are absent from the supplied evidence, so only limited static verifiability is justified.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →