Claude Skill Creator FAQs

Question 1

What is the difference between a Claude Skill and a slash command?

Accepted Answer

Slash commands require manual invocation (e.g., /command), whereas skills are directory-based capabilities that Claude automatically discovers and applies based on the conversation context.

Question 2

How does Claude know when to trigger a specific skill?

Accepted Answer

Claude uses the 'description' field in the SKILL.md frontmatter. A well-written description includes specific actions and trigger keywords that help the model decide when the skill is relevant.

Question 3

What is 'forked context' in Claude Code skills?

Accepted Answer

Setting context to 'fork' allows a skill to run as an isolated sub-agent, which is ideal for long-running or complex operations that shouldn't clutter the main chat history.

Question 4

Can I restrict what commands a skill can execute?

Accepted Answer

Yes, you can use the 'allowed-tools' field in the frontmatter to define specific permissions, such as limiting Bash commands to specific patterns like 'git:*' or 'python:*'.

Question 5

Where does this skill store newly created capabilities?

Accepted Answer

Skills can be created in the global personal directory (~/.config/claude/skills/) for use in all projects, or in the project-specific directory (.claude/skills/) for local use.

Claude Skill Creator

主要功能

使用场景

Claude Skill Creator

主要功能

使用场景