6 Claude Code Slash

06 — Slash command + subagent + skill, all inside Claude Code

Claude Code example: a slash command orchestrates a research subagent and a formatting skill to produce a travel trip brief.

**Tags:** `claude-code` `anthropic` `claude` `ai-agents` `subagents` `slash-commands` `agentic-workflows` `llm` `automation` `prompt-engineering`

The whole thing runs **in the Claude Code app**, driven by **slash syntax**. Nothing to compile or host — Claude Code auto-discovers the files under `.claude/` and wires them together.

06-claude-code-slash/
└── .claude/
    ├── commands/trip-plan.md        ← slash command:  /trip-plan   (entry point)
    ├── agents/trip-researcher.md    ← subagent:       does the research
    └── skills/trip-brief/SKILL.md   ← skill:          formats the output

How the three pieces connect

1. /trip-plan — a slash command (.claude/commands/trip-plan.md). $ARGUMENTS receives whatever you type after the command. This is the entry point and the orchestration prompt.
2. It delegates research to the trip-researcher subagent (.claude/agents/trip-researcher.md) — its own context window, its own allowed tools (WebSearch, WebFetch), running on a cheaper model.
3. It formats the researcher's facts using the trip-brief skill (.claude/skills/trip-brief/SKILL.md), which Claude loads on demand.

So: **slash command orchestrates → subagent researches → skill formats.** All three are just Markdown files; Claude Code is the runtime.

Run it

cd 06-claude-code-slash
claude                 # open the Claude Code app in this directory

Then type:

/trip-plan Lisbon 2026-08-15

Claude Code kicks off the command, spawns the `trip-researcher` subagent to gather weather + flight facts, then hands them to the `trip-brief` skill and prints the finished brief.

• Type / to see trip-plan in the slash-command list.
• These live in the project .claude/. Move them to `~/.c