What does this skill do, and when should you use it?
This skill supports developers who create or improve skills for Claude Code plugins. It explains SKILL.md metadata and structure, along with optional scripts, references, examples, and assets. Its main focus is trigger descriptions, progressive disclosure, imperative writing, and validation. It is intended for plugin-skill development rather than general coding or code-generation work.
Provides a staged workflow for skill creation: understand use cases, plan reusable resources, create the plugin skill directory, edit SKILL.md, add references, examples, and scripts, then check structure, frontmatter, trigger phrases, writing style, referenced files, and scripts. It supplies mkdir and touch examples for creating files and gives the cc --plugin-dir command for local plugin testing. It also recommends asking the skill-reviewer agent to review skill quality.
- A plugin developer needs to create a new skill containing SKILL.md.
- A team needs to add scripts, references, or examples to an existing skill.
- A maintainer wants stronger trigger phrases so the skill activates for the right requests.
- A developer needs to split lengthy instructions into progressively loaded reference material.
- A maintainer wants to validate structure, writing style, and local loading behavior before release.
How do you install this skill?
- This assessment is based only on the supplied static files; no commands, tests, or installation flows were executed.
- The skill does not clearly define licensing, maintenance ownership, update mechanisms, rollback, or sensitive-data boundaries.
- Some plugin behavior, directory contents, and validation assumptions depend on repository context outside the assessed skill evidence.
- Chinese-language interaction and mainland-China network compatibility are not documented.
- Shell / CLI
- Local filesystem
Claude Code plugin environment
The README documents installation of Claude Code as a whole through the macOS/Linux curl installer, Homebrew, Windows PowerShell, or WinGet; npm installation is marked deprecated. It provides no standalone installation command for this skill. The skill is located at plugins/plugin-dev/skills/skill-development/SKILL.md and is used as part of the relevant Claude Code plugin repository.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/anthropics/claude-code.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/plugins/plugin-dev/skills/skill-development" ~/.claude/skills/
rm -rf "$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 skill,
- add a skill to plugin,
- improve skill description,
- organize skill content.
In Claude Code, make a request such as “create a skill,” “add a skill to plugin,” “improve skill description,” or “organize skill content.” For a new plugin skill, create plugin-name/skills/skill-name/ and add SKILL.md as described. For local testing, the guide provides cc --plugin-dir /path/to/plugin.
What are this skill's strengths and limitations?
- Covers the full workflow from understanding use cases through iteration.
- Defines concrete standards for trigger descriptions, imperative prose, and progressive disclosure.
- Includes practical checks for structure, references, validation, and local testing.
- Explains Claude Code plugin directories and automatic skill discovery.
- It is centered on Claude Code plugins, so directories, commands, and platform assumptions require adjustment elsewhere.
- No standalone installation or packaging workflow is documented.
- The scripts, references, and examples shown are guidance, not evidence of bundled resources in this skill.
- The repository license is listed as unknown in the supplied source.
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 Skill Development this page ✓ Anthropic · Official | 46 · Use with care | ★ 151k | 11d ago | — |
| Claude Code Plugin Skill Development ✓ Anthropic · Official | 51 · 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 |
| Project Plugin Settings Pattern ✓ Anthropic · Official | 49 · Use with care | ★ 38k | 11d ago | Apache-2.0 |
| Claude Code Command Builder ✓ Anthropic · Official | 46 · Use with care | ★ 38k | 11d ago | Apache-2.0 |
How did FollowSkills review this skill?
The material provides guidance for skill authoring and validation and shows no malware, credential theft, covert exfiltration, or destructive default. It lacks explicit user confirmation, sensitive-data handling, external-effect controls, rollback guidance, and dependency-security treatment, so points are deducted.
The structure, trigger guidance, and validation checklist are broadly consistent and include directory, command, and failure-check guidance. However, there are no executed test results, abnormal-input cases, or diagnosable failure outputs. The static-review ceiling supports 7 points.
The audience and use case—Claude Code plugin skill development—are clear, with concrete trigger phrases and a defined workflow. Non-fit boundaries, Chinese-language support, environment compatibility, and mainland-China network reachability are not addressed, so points are deducted.
The document has clear information architecture and covers progressive disclosure, directory conventions, examples, version metadata, and resource references. The selected skill does not state a license, changelog, maintenance owner, or explicit update path, and some examples rely on surrounding repository context, so points are deducted.
The workflow from use-case discovery through resource planning, structure creation, editing, validation, and iteration is actionable and covers the core task. However, no representative output or execution evidence verifies the key paths, and some guidance requires additional repository context. The static ceiling supports 6 points.
The supplied source is auditable and contains explicit rules, checklists, and command examples. It provides no skill-specific test suite, CI coverage, or third-party execution evidence, and does not systematically separate facts from recommendations, so 4 points are awarded.
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 →