
Explain Code
Purpose → flow → gotchas at the altitude you asked for — reads the real code, corrects misleading names, skips line-by-line noise.
v1.0.0 · ~543 tokens · ⬇ 0 · Updated July 6, 2026
What it does
Explains code at the right altitude (repo overview vs function detail vs file walk-through): purpose before mechanism, the control/data flow, and the highest-value part — the gotchas a newcomer can't see (implicit assumptions, hidden dependencies, historical weirdness). Reads the actual code, corrects misleading names, and traces a concrete example. Skips restating obvious lines.
Example uses
Onboard to an unfamiliar repo
It is your first day on a codebase and you need the map before you start opening random files.
I just cloned this repo. Give me the big picture: what it does, the entry points, how the main pieces fit together, and the architecture choices I should know about. Don't dive into individual files yet — I need the map first.Decode a legacy function
A critical function has no comments, a misleading name, and the person who wrote it left the company.
Walk me through the processQueue() function in src/workers/queue.js. Give me its job in one sentence, then inputs, side effects, and the gotchas — especially anything that breaks if I change the retry logic. If the name is misleading, say so and tell me what it actually does.Trace a request through the code
You need to understand a data flow end to end before you touch any part of it.
Trace what happens in this codebase when a user submits the checkout form: where the request enters, how the payment data transforms, and where it exits. Focus on the flow and the key decisions, not line-by-line narration, and flag anything that looks like a landmine while you're in there.Install
# 1. Create the skill folder in your Claude setup mkdir -p ~/.claude/skills/explain-code # 2. Download SKILL.md into it (or move the file you just downloaded) # → ~/.claude/skills/explain-code/SKILL.md # 3. Claude Code auto-discovers it on next launch.
Inside the skill
---
name: explain-code
description: Explain what a piece of code, a function, or a codebase actually does — clearly, at the right depth. Use when onboarding to unfamiliar code, understanding a legacy function, or asked "explain this code", "what does this do", "walk me through this file/repo". Explains the WHY and the flow, not line-by-line noise.
---
# Explain Code
Good code explanation gives understanding you can act on — the purpose, the flow, and the
gotchas — not a translation of each line into prose. Read the real code; don't guess from names.
## Explain at the right altitude (match the ask)
- **"What is this repo?"** → the big picture first: what it does, entry points, how the main
pieces fit, the tech/architecture. Don't dive into a random file.
- **"What does this function do?"** → its job in one sentence, then inputs → what it does →
outputs/side effects, then the non-obvious parts (edge cases, why that weird bit exists).
- **"Walk me through this file"** → the flow: what calls what, the data's journey, the key
decisions — not every line.
## What actually helps
- **Purpose before mechanism**: WHY this exists / what problem it solves, then HOW. Mechanism
without purpose is just narration.
- **The flow / data journey**: where input enters, how it transforms, where it exits. Control
flow and data flow are what people actually need to hold in their head.
- **The gotchas**: the non-obvious dependency, the implicit assumption, the "if you change this,
that breaks", the historical weirdness. This is the highest-value part — it's what a newcomer can't see.
- **Concrete example**: trace one real input through it. An example teaches faster than abstraction.
## Rules
- Don't restate obvious lines ("this loop iterates over the users") — that adds nothing.
- Don't guess. If a name is misleading or intent is unclear, read what it actually does and
say so ("despite the name, this doesn't cache — it…") — never invent a rationale.
## Output
An explanation at the altitude asked: purpose → flow → gotchas, with a concrete example if it
helps. Flag anything that looks like a bug or a landmine you noticed while reading.
Changelog
- v1.0.02026-07-03Initial clean-room write.
Frequently asked questions
Is Explain Code free?
Yes. Explain Code is free to download and MIT-licensed.
Where do I install Explain Code?
Place the SKILL.md file in ~/.claude/skills/explain-code/ and Claude Code auto-discovers it on next launch.
How many tokens does Explain Code use?
About 543 tokens — it is designed to be token-lean.

We're developers and SaaS builders who use these tools daily in production. Every review comes from hands-on experience building real products — DealPropFirm, ThePlanetIndicator, PropFirmsCodes, and many more. We don't just review tools — we build and ship with them every day.
Written and tested by developers who build with these tools daily.