agentleFS
arimxyerarimxyer/toolchest/.claude/skills/scaffold-plugin/SKILL.md

scaffold-plugin

Skill1 starsChanged 5 months agoApache-2.0View on GitHub

Scaffold a new plugin in the toolchest marketplace. Creates plugins/<name>/ with .claude-plugin/plugin.json, README.md, and skills/<skill>/SKILL.md, then patches the root marketplace.json and the root README plugin table. Use when the user asks to add, scaffold, or create a new plugin.

4 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions4 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 4

    1. TrogonStack/agentskills★ 11same name
    2. helmut-hoffer-von-ankershoffen/heinzel★ 2same name
    3. g-bastianelli/nuthouse★ 1same name
    4. farzanmrz/meta-dev★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository2 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

9 rules. 0 are common, 4 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Add a new plugin to this marketplace, keeping the three sources of truth in sync.

    line 11
  2. Very specific:

    If $1 is missing, ask the user.

    line 18
  3. Very specific:

    Leave optional frontmatter (argument-hint, allowed-tools, metadata) off the starter SKILL.md.

    line 104
  4. Very specific:

    If a skill ships supporting files (templates, examples), put them in skills/<skill>/references/ next to that skill — not at the plugin root …

    line 106
  5. Very specific:

    Use the Edit tool with a unique anchor (e.g., the closing ] of plugins[] plus the preceding }) to insert.

    line 125
  6. Distinctive:

    After writing all files, run:

    line 133
  7. Distinctive:

    Don't commit the changes — leave that to the user.

    line 144
  8. Very specific:

    Don't run /check-marketplace-sync automatically.

    line 145
  9. Distinctive:

    Don't fabricate a real implementation for the first skill — the skeleton above is intentional.

    line 146

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

scaffold-plugin

Add a new plugin to this marketplace, keeping the three sources of truth in sync.

Inputs

  • $1 — plugin name (kebab-case, required). Becomes the directory name and the name field everywhere.
  • $2 — first skill name (optional, kebab-case). Defaults to the plugin name.

If $1 is missing, ask the user. Do not invent a name.

Before doing anything

Gather these from the user in a single AskUserQuestion call (don't echo back the plugin name — ask only what you don't have):

  1. One-line description for the plugin. Will appear in plugin.json, marketplace.json, and the root README row.
  2. First skill description (one line). Will appear in the first SKILL.md frontmatter and the README row.

Then propose the full file plan in chat before writing — list every file you'll create and every file you'll modify with a one-line summary. Wait for approval.

Plugin shape: skills vs LSP vs agents

This template scaffolds a skill-based plugin. Two other shapes exist; the three-place sync and metadata conventions below apply unchanged to all of them.

  • LSP-server plugin (like ty-lsp): skip skills/<skill>/SKILL.md and the first-skill questions entirely, and instead create .lsp.json at the plugin root (server-name → {command, args, extensionToLanguage}); the root README table's Skill column gets —.
  • Agent-bundling plugin (like writers-room): add agents/<agent>.md files at the plugin root, one per agent — frontmatter needs name and description; tools is optional (note: agents use tools, skills use allowed-tools). Agents are auto-discovered (no plugin.json field) and register as <plugin>:<agent>. If one agent spawns siblings, its Agent(...) allowlist must use the namespaced form — Agent(<plugin>:<agent>, …); bare names break delegation silently. Adding or renaming an agent means updating every sibling allowlist that names it. A skill that needs the user (interview, confirmation gates) stays main-thread — subagents have no AskUserQuestion.

File plan

Create:

  • plugins/<plugin>/.claude-plugin/plugin.json
  • plugins/<plugin>/README.md
  • plugins/<plugin>/skills/<skill>/SKILL.md

Modify:

  • .claude-plugin/marketplace.json — append plugin entry to plugins[]
  • README.md — append row to "Available Plugins" table

Templates

Base every field on existing files. The owner string is arimxyer and the license is Apache-2.0 — do not substitute the local git user.

plugins/<plugin>/.claude-plugin/plugin.json

{
  "name": "<plugin>",
  "description": "<one-line description>",
  "version": "1.0.0",
  "author": {
    "name": "arimxyer"
  },
  "repository": "https://github.com/arimxyer/toolchest",
  "license": "Apache-2.0"
}

plugins/<plugin>/README.md

Minimal — mirror the structure of plugins/sudo-askpass/README.md:

# <Title-Case Plugin Name>

<One-paragraph description of what the plugin does.>

## What It Does

- <bullet>
- <bullet>

## Usage

It can also be invoked directly:

```text
/<plugin>:<skill> <args>

### `plugins/<plugin>/skills/<skill>/SKILL.md`

```markdown
---
name: <skill>
description: <one-line skill description>
---

# <skill>

<Body of the skill — what it does and when to use it.>

Leave optional frontmatter (argument-hint, allowed-tools, metadata) off the starter SKILL.md. Frontmatter style varies across this repo's plugins; the user can add fields when the skill's shape is clear.

If a skill ships supporting files (templates, examples), put them in skills/<skill>/references/ next to that skill — not at the plugin root (see writers-room's MEMO.template.md / BRAND.template.md).

.claude-plugin/marketplace.json

Append to the plugins array, matching the existing object shape exactly:

{
  "name": "<plugin>",
  "source": "./plugins/<plugin>",
  "description": "<one-line description>",
  "version": "1.0.0",
  "author": {
    "name": "arimxyer"
  },
  "license": "Apache-2.0"
}

Use the Edit tool with a unique anchor (e.g., the closing ] of plugins[] plus the preceding }) to insert. After editing, verify by reading the file back. The prettier hook will reformat it on save — that's expected.

README.md

Append a row to the "Available Plugins" table. Match the existing column order: | <plugin> | /<plugin>:<skill> | <description> |.

Verification

After writing all files, run:

jq -e . .claude-plugin/marketplace.json > /dev/null
jq -e . plugins/<plugin>/.claude-plugin/plugin.json > /dev/null

Both should exit 0. Then list what was created and ask the user to review.

Out of scope

  • Don't commit the changes — leave that to the user.
  • Don't run /check-marketplace-sync automatically; offer it as a follow-up if useful.
  • Don't fabricate a real implementation for the first skill — the skeleton above is intentional. The user will fill in the body.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.