# Zoom Out

> Instructs the agent to map the call chain and boundaries before making changes, providing broader context or a higher-level perspective. Use when unfamiliar with a section of code or needing to understand how it fits into the bigger picture.

- Skill: `tyecode/zoom-out` (Agent Skill)
- Install (CLI): `npx skillmds@latest add tyecode/zoom-out`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tyecode/zoom-out/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: tyecode (https://skillmd.com/u/tyecode)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/tyecode/zoom-out

---


# Skill: zoom-out

## When to Use This Skill

Use this skill when:
- You're about to edit code you haven't read the full context of
- Something is behaving unexpectedly and you don't know why
- You've been deep in a file and lost track of the bigger picture

## The Process

Do this before making changes, not after:

1. **Find the entry point** — where does execution start for this feature? Trace from the route, CLI command, or event trigger.
2. **Map the call chain** — who calls what. Write it out as a list: `A → B → C → D`.
3. **Identify the boundaries** — where does this code touch external systems? (DB, API, filesystem, queue)
4. **Read the types** — what data flows through? Types tell you intent faster than implementation.
5. **Check the tests** — existing tests describe expected behavior. Read them before reading implementation.

## What to Produce

Before diving in, write a short summary:
- What this code does in one sentence
- The call chain from entry to the code in question
- What you expected vs. what you found (if something surprised you)

This takes 2 minutes and prevents 30-minute wrong turns.

## When You're Truly Lost

If you can't map the flow after reading the code, say so explicitly:
> "I can see what this code does line by line, but I don't understand why it exists or what calls it. I need more context before making changes."

Then ask the user to explain it or point to the entry point. Never make changes to code you don't understand.