[IMPORTANT] Use TaskCreate to break ALL work into small tasks BEFORE starting — including tasks for each file read. This prevents context loss from long files. For simple tasks, AI MUST ask user whether to skip.
Quick Summary
Goal: Add new reference files or scripts to an existing skill directory.
Workflow:
-
Identify — Determine target skill and required additions
-
Create — Add reference/script files following skill-creator patterns
-
Validate — Ensure files follow progressive disclosure (<100 lines each)
Key Rules:
-
Delegates to skill-creator and claude-code skills for guidance
-
Reference files must be under 100 lines (progressive disclosure)
-
Scripts must have tests and respect .env loading order
Be skeptical. Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence percentages (Idea should be more than 80%).
Think harder. Use skill-creator and claude-code skills. Use docs-seeker skills to search for documentation if needed.
Arguments
$1: skill name (required, default: "") $2: reference or script prompt (required, default: "") If $1 or $2 is not provided, ask the user to provide it.
Your mission
Add new reference files or scripts to a skill at .claude/skills/$1 directory.
Requirements
Rules of Skill Creation:
Base on the requirements:
-
Always keep in mind that SKILL.md and reference files should be token consumption efficient, so that progressive disclosure can be leveraged at best.
-
SKILL.md is always short and concise, straight to the point, treat it as a quick reference guide.
-
If you're given nothing, use AskUserQuestion tool for clarifications and researcher subagent to research about the topic.
-
If you're given an URL, it's documentation page, use Explore subagent to explore every internal link and report back to main agent, don't skip any link.
-
If you receive a lot of URLs, use multiple Explore subagents to explore them in parallel, then report back to main agent.
-
If you receive a lot of files, use multiple Explore subagents to explore them in parallel, then report back to main agent.
-
If you're given a Github URL, use repomix command to summarize (install it if needed) and spawn multiple Explore subagents to explore it in parallel, then report back to main agent.
IMPORTANT:
-
Skills are not documentation, they are practical instructions for Claude Code to use the tools, packages, plugins or APIs to achieve the tasks.
-
Each skill teaches Claude how to perform a specific development task, not what a tool does.
-
Claude Code can activate multiple skills automatically to achieve the user's request.
IMPORTANT Task Planning Notes (MUST FOLLOW)
-
Always plan and break work into many small todo tasks
-
Always add a final review todo task to verify work quality and identify fixes/enhancements