Skill Maker

Skill Maker

Protocols

Read protocol: ${PG_PROTOCOLS}/ux-protocol.md (use the read_file tool before continuing).

Read protocol: ${PG_PROTOCOLS}/input-validation.md (use the read_file tool before continuing).

Read protocol: ${PG_PROTOCOLS}/tool-efficiency.md (use the read_file tool before continuing).

Read protocol: ${PG_PROTOCOLS}/visual-identity.md (use the read_file tool before continuing).

Read protocol: ${PG_PROTOCOLS}/freshness-protocol.md (use the read_file tool before continuing).

Read protocol: ${PG_PROTOCOLS}/receipt-protocol.md (use the read_file tool before continuing).

Read protocol: ${PG_PROTOCOLS}/boundary-safety.md (use the read_file tool before continuing).

Read protocol: ${PG_PROTOCOLS}/conflict-resolution.md (use the read_file tool before continuing).

Run shell command before continuing: cat .production-grade.yaml 2>/dev/null || echo "No config — using defaults" (use the execute_shell_command tool).

Fallback (if protocols not loaded): Use AskUserQuestion with options (never open-ended), "Chat about this" last, recommended first. Work continuously. Print progress constantly. Validate inputs before starting — classify missing as Critical (stop), Degraded (warn, continue partial), or Optional (skip silently). Use parallel tool calls for independent reads. Use smart_outline before full Read.

Progress Output

Follow Claude-Production-Grade-Suite/.protocols/visual-identity.md. Print structured progress throughout execution.

Skill header (print on start):

━━━ Skill Maker ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Phase progress (print during execution):

  [1/3] Pattern Analysis
    ✓ {N} recurring patterns identified
    ⧖ analyzing workflow structure...
    ○ skill generation
    ○ installation

  [2/3] Skill Generation
    ✓ {N} custom skills drafted
    ⧖ writing SKILL.md files...
    ○ installation

  [3/3] Installation
    ✓ {N} skills installed to .claude/skills/

Completion summary (print on finish — MUST include concrete numbers):

✓ Skill Maker    {N} project-specific skills created    ⏱ Xm Ys

Overview

End-to-end skill and plugin creation pipeline. Interviews the user on what the skill should do, writes the SKILL.md, packages it as a Claude Code plugin, creates a GitHub repo, and adds it to the user's marketplace — all in one flow.

Config Paths

Read .production-grade.yaml at startup if available. Skill-maker is mostly self-contained and does not depend on project-level path overrides.

When to Use

• User asks to create a new skill or plugin
• User describes a reusable workflow that should be a skill
• User says "make a skill", "build a plugin", "I need a skill for..."
• NOT for: editing existing skills (just edit the file directly)

Process Flow

digraph skill_maker {
    rankdir=TB;

    "Skill idea received" [shape=doublecircle];
    "Phase 1: Interview" [shape=box];
    "Phase 2: Write SKILL.md" [shape=box];
    "User approves?" [shape=diamond];
    "Phase 3: Package as plugin" [shape=box];
    "Phase 4: Create repo & push" [shape=box];
    "Phase 5: Add to marketplace" [shape=box];
    "Done" [shape=doublecircle];

    "Skill idea received" -> "Phase 1: Interview";
    "Phase 1: Interview" -> "Phase 2: Write SKILL.md";
    "Phase 2: Write SKILL.md" -> "User approves?";
    "User approves?" -> "Phase 2: Write SKILL.md" [label="revise"];
    "User approves?" -> "Phase 3: Package as plugin" [label="approved"];
    "Phase 3: Package as plugin" -> "Phase 4: Create repo & push";
    "Phase 4: Create repo & push" -> "Phase 5: Add to marketplace";
    "Phase 5: Add to marketplace" -> "Done";
}

Phase 1: Interview (Quick)

Engagement mode determines interview depth:

• Express: Infer all answers from the project context and agent prompts. Skip interview entirely. Report inferences.
• Standard: Ask 1-2 key questions (purpose + trigger conditions). Auto-resolve the rest.
• Thorough/Meticulous: Ask 3-4 questions using AskUserQuestion, one at a time:

1. What does this skill do? — Core purpose in one sentence
2. When should it trigger? — Specific words, patterns, or situations
3. What's the workflow? — Steps the skill should follow (linear, loop, decision tree?)
4. Skill type? — Options: Technique (steps to follow), Pattern (mental model), Reference (docs/API guide), Workflow (multi-phase process)

Phase 2: Write SKILL.md

Follow these rules from the writing-skills methodology:

Frontmatter:

• name: kebab-case, letters/numbers/hyphens only
• description: Start with "Use when...", max 500 chars, triggering conditions only — NEVER summarize the workflow

Structure:

---
name: skill-name
description: Use when [triggering conditions]
---

# Skill Name

## Overview
Core principle in 1-2 sentences.

## When to Use
Bullet list with symptoms and use cases.

## Process Flow (if multi-step)
Small inline dot flowchart for non-obvious decisions.

## [Core Content]
Steps, patterns, or reference material.

## Common Mistakes
Table of mistake -> fix pairs.

Quality rules:

• One excellent example beats many mediocre ones
• Flowcharts ONLY for non-obvious decision points
• Keep under 500 words for standard skills
• Use active voice, verb-first naming
• Include keywords for discoverability (error messages, symptoms, tool names)

Skill review (mode-aware): Express — proceed to packaging, report skill summary. Standard — present brief summary for approval. Thorough/Meticulous — present full SKILL.md for detailed review via AskUserQuestion.

Phase 3: Package as Plugin

Create the plugin directory structure:

<skill-name>-plugin/
├── .claude-plugin/
│   └── plugin.json
├── skills/
│   └── <skill-name>/
│       └── SKILL.md
└── README.md

plugin.json template:

{
  "name": "<skill-name>",
  "description": "<one-line description>",
  "version": "1.0.0",
  "author": {
    "name": "<from git config or ask>"
  },
  "license": "MIT",
  "keywords": ["<relevant>", "<tags>"]
}

README.md template:

# <Skill Name> Plugin for Claude Code

<description>

## Installation

### Via Marketplace
/plugin marketplace add nagisanzenin/claude-code-plugins
/plugin install <skill-name>@nagisanzenin

### Load Directly
claude --plugin-dir /path/to/<skill-name>-plugin

## Usage
<trigger description and examples>

## License
MIT

Phase 4: Create Repo & Push

1. git init in the plugin directory
2. git add -A && git commit -m "Initial release: <skill-name> plugin v1.0.0"
3. gh repo create <skill-name>-plugin --public --source . --push
4. If gh auth fails, guide user through gh auth login

Phase 5: Add to Marketplace

1. Read the user's marketplace repo (default: nagisanzenin/claude-code-plugins)
2. Clone or locate the marketplace locally
3. Add new plugin entry to .claude-plugin/marketplace.json:

   {
     "name": "<skill-name>",
     "source": {
       "source": "github",
       "repo": "nagisanzenin/<skill-name>-plugin"
     },
     "description": "<description>",
     "version": "1.0.0"
   }
   

4. Update README.md plugin table
5. Commit and push marketplace repo
6. Report final install command: /plugin install <skill-name>@nagisanzenin

Marketplace Config

Default marketplace repo: nagisanzenin/claude-code-plugins Default marketplace local path: ~/nagisanzenin-plugins Default plugin location: ~/<skill-name>-plugin

If the user has a different marketplace, ask which one to use.

Common Mistakes

This skill body has been adapted for QwenPaw. Differences vs the upstream Claude Code plugin to be aware of:

• No AskUserQuestion tool. When this skill says to surface a decision, render numbered options as plain Markdown and ask the user to type the option name. Parse free-text replies leniently.
• No Skill tool. Phase transitions happen in-line: read the next sub-skill body via read_file from the workspace skills/ dir.
• No subagent spawn. v0.1 is a single-agent flow. If the methodology says "delegate to specialist X", invoke X by reading its SKILL.md from skills/<name>/SKILL.md and following its instructions yourself.
• No TaskCreate/TaskList. Track progress by writing receipts to Claude-Production-Grade-Suite/.orchestrator/receipts/<task>-<role>.json and emitting a one-line status update in chat after each phase.
• WebSearch is tavily_search. Requires TAVILY_API_KEY. If unset, skip the Freshness Protocol and note it.