saadshahd

/generate — Development skill for Claude Code

Development community intermediate

Create production-ready .md files for the moo marketplace.

How to install /generate

This entry records only its repository, not the path inside it, so there is no exact command to give. Open saadshahd/moo.md and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What /generate does

Create production-ready `.md` files for the moo marketplace. **Internally applies:** `/prompt` techniques to all generated content.

Alternatives in Development

  • Gemini CLI Foundations — This repository is a comprehensive skills library for Gemini CLI - reusable, production-ready skill packages t 5.3k ★
  • Claude Hooks Mastery — 3,300+ Complete mastery guide for Claude Code hooks with production-ready scripts 3.4k ★
  • Axiom — Battle-tested Claude Code skills for modern xOS (iOS, iPadOS, watchOS, tvOS) development with 13 production-re 2.6k ★

README


description: Generate command/skill/agent files for moo ecosystem. Applies /prompt techniques automatically.

/generate

Create production-ready `.md` files for the moo marketplace.

**Internally applies:** `/prompt` techniques to all generated content.

Inputs

  • $1 — Type (command | skill | agent) or infer
  • $2 — Task description

Ecosystem Structure

Where Files Live

Type Project-level Plugin-level
Command .claude/commands/X.md [plugin]/commands/X.md
Skill — [plugin]/skills/[name]/SKILL.md
Agent .claude/agents/X.md [plugin]/agents/X.md
Ref — [plugin]/skills/[name]/references/

Plugin Architecture

[plugin]/
├── .claude-plugin/plugin.json    # name, version, description
├── skills/[name]/
│   ├── SKILL.md                  # Router, auto-triggers
│   └── references/*.md           # Supporting content
└── commands/*.md                 # User-invoked

Frontmatter Rules

**Commands/Agents:**

---
description: Single line. Max 1024 chars. No | or >.
---

**Skills:**

---
name: kebab-case
description: Single line. Trigger + action.
version: 0.0.1
---

Process

Phase 1: Parse

  • What — Core function (analyze, generate, debug, search, evaluate)
  • Where — Project-level or plugin? Which plugin?
  • Type — Command (explicit), Skill (background), Agent (delegated)

Phase 2: Confidence Gate

Confidence Action
< 70% Ask ONE question, STOP
70-85% Add ## Assumptions
≥ 85% Generate directly

Phase 3: Type Detection

Signal Type
"when I run", "command for" Command
"always apply", "principles", background Skill
"search", "find", "analyze", delegatable Agent

Default → Command.

Phase 4: Generate

Read `/prompt` and apply its techniques, then use ecosystem template:

Command Structure

# /[name]
[Purpose]
## When to Use
## Inputs
## Process (numbered steps)
## Output Format
## Examples

Skill Structure

# [Name]
[Philosophy]
## When This Applies
## Core Principles
## Process
## Anti-Patterns

Agent Structure

# [Name]
[Role + constraint]
## Delegation Trigger
## Strategy
## Output Format

Conventions

  • Replace "think" → "consider/evaluate/reason"
  • Use "Use when..." not "MUST/CRITICAL"
  • Story points, never time estimates
  • Quality footer for non-trivial outputs
  • Single-level references (no A→B→C chains)

Post-Generation

  1. State full file path
  2. Verify frontmatter (single-line, no multi-li