Guide users through creating modular AI skills with SKILL.md structure.
npx skills add https://github.com/bmad-labs/skills --skill skill-creator-bmad-labs
Please help me install this Agent Skill. Skill: skill-creator Source: https://github.com/bmad-labs/skills/tree/main/skills/skill-creator Command: npx skills add https://github.com/bmad-labs/skills --skill skill-creator-bmad-labs
This Skill provides a comprehensive guide and process for creating effective, token-efficient, and well-structured AI skills, ensuring Claude can reliably extend its capabilities.
SKILL.md, scripts/, references/, and assets/ for modular skill design.references/ directory.Use the skill-creator skill to learn how to create a new skill for Claude.
High-intent search queries and answers about installing and using this skill.
To create new AI skills for Claude, follow a structured development process emphasizing efficiency and modularity. This involves understanding skill anatomy, applying core principles like conciseness and progressive disclosure, and organizing scripts and references effectively.
Progressive disclosure in AI skill development is a core principle for managing context window usage. It involves layered information loading, ensuring Claude receives only necessary data at each step to maintain token efficiency and reliable capability extension.
Organize scripts and references for a Claude skill using modular directories. Place executable code in scripts/ and supporting documentation like API guides in references/, ensuring the SKILL.md file orchestrates these components cleanly.
The best way to structure documentation for AI workflow automation is through modular skill design. Use a SKILL.md file for core instructions and store detailed API documentation within the references/ directory to enable efficient, context-aware retrieval.
AI skills fail to reliably extend capabilities when they lack conciseness and proper modular structure. Without a structured development process and organized assets, context windows become overloaded, preventing Claude from interpreting instructions accurately.
ralphjohn29Create, test, and benchmark custom AI skills with SKILL.md files.
Dev10x-GuruAudit Claude context window usage against documented budgets and suggest pruning actions.
hongmaple0820Structure and load AI agent context within token budgets.
SJTU-IPADSDesign, analyze, and optimize prompts with evaluation frameworks and structured outputs.