skill-creator

Skill Creator

Setup

Before starting: check .handoff/sessions/ for active sessions, read context status.yaml, run git status. Follow .rules/universal.md (Plan -> Approve -> Execute).

Create effective skills that extend Claude's capabilities for the POS workflow.

Core Principles

Context is Precious

The context window is shared with system prompts, conversation history, and user requests. Only include information Claude doesn't already have. Challenge each piece: "Does this justify its token cost?"

Graduated Information Loading

1. Metadata (~100 words) - Always in context (name + description)
2. SKILL.md body (<500 lines) - Loaded when skill triggers
3. Bundled resources (unlimited) - Loaded as needed

Appropriate Constraints

• High freedom: Multiple valid approaches, use text guidance
• Medium freedom: Preferred patterns exist, use pseudocode
• Low freedom: Fragile operations, use specific scripts

Skill Structure

skill-name/
├── SKILL.md (required)
│   ├── YAML frontmatter (name, description, allowed-tools)
│   └── Markdown instructions (<500 lines)
└── Bundled Resources (optional)
    ├── scripts/      - Executable code for deterministic tasks
    ├── references/   - Documentation loaded contextually
    └── assets/       - Output resources (templates, images)

Creation Process

Step 1: Understand the Skill

Ask clarifying questions:

• "What functionality should this skill support?"
• "Give examples of how this skill would be used"
• "What should trigger this skill?"

Step 2: Plan Resources

For each use case, identify:

• Scripts for repetitive code
• References for documentation
• Assets for templates/output

Step 3: Initialize Structure

Create the skill directory:

mkdir -p .claude/skills/{skill-name}/{scripts,references,assets}

Step 4: Write SKILL.md

Frontmatter (CRITICAL)

---
name: skill-name
description: What it does AND when to use it. Include trigger phrases.
allowed-tools: Read, Write, Edit, Glob, Grep, Bash, Agent
---