← ClaudeAtlas

core-createlisted

Create new Claude Code skills interactively. Use with /core-create.
AirMile/claude-config · ★ 0 · AI & Automation · score 75
Install: claude install-skill AirMile/claude-config
# Create Create new skills through a streamlined process: gather requirements → plan → create. **Trigger**: `/core-create` or `/core-create [name]` ## Skill Spec Quick Reference ### Frontmatter Required: `name` (kebab-case, matches folder) and `description` (WHAT + WHEN pattern with trigger phrases, under 1024 chars). Optional Claude Code fields: `argument-hint`, `disable-model-invocation`, `user-invocable`, `allowed-tools`, `context: fork`, `agent`. Recommended: `metadata` with author (`your-username`), version, category (`core|dev|frontend|game|project|story|team|thinking`). ### Description Pattern `[What it does] + [When to use it / trigger phrases]` Good: `"Analyze staged git changes and generate conventional commit messages. Use with /core-commit. Detects rebase/merge state, validates changes."` Bad: `"Helps with projects."` — too vague, no triggers. ### Cross-Skill References When referencing other skills, ALWAYS use hyphen notation matching the directory name: - Correct: `/dev-build`, `/game-verify`, `/frontend-design` - Wrong: `/dev:build`, `/game:test`, `/frontend:compose` The colon notation does NOT match skill directory names and breaks slash command resolution. ### Security Rules - No XML angle brackets (< >) in frontmatter - No "claude" or "anthropic" in skill name - Description under 1024 characters ### Substitutions - `$ARGUMENTS` / `$ARGUMENTS[N]` / `$N` — argument injection - `` !`command` `` — dynamic context injection (shell commands run