How do I install the AI Explain skill?

Run npx skillmds add majiayu000/ai-explain in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is AI Explain safe to use?

This skill has not completed SkillMD's automated safety review yet. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with AI Explain?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is AI Explain free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published AI Explain?

majiayu000 (@majiayu000) published this skill. Their other Agent Skills are listed on their SkillMD profile.

AI Explain by majiayu000 — Agent Skill

Name: AI Explain
Rating: 4.2 (5 reviews)
Author: majiayu000

File contents

Explain

Engineer-grade technical explanations of code, concepts, patterns, and architecture. 3-tier depth control scales detail to what the developer needs. Anchored in the actual codebase with file:line references, ASCII diagrams, and execution traces.

When to Use

"How does this work?", "What is this?", "Why does this do X?", "Trace this."
NOT for generating documentation -- use /ai-write.
NOT for writing/fixing code -- use ai-build agent or /ai-debug.

Process

1. Identify subject

Classify into: Code, Concept, Pattern, Architecture, Error, or Difference. If ambiguous, ask ONE clarifying question.

2. Search codebase

Use Grep/Glob to find real instances. Codebase examples with file:line references are primary evidence. If not found, use generic example in project's stack and note it.

3. Select depth

Depth	Trigger cues	Sections
Brief	"TL;DR", "brief", "short"	Summary + Walkthrough (3-5 steps)
Standard	General question (DEFAULT)	Summary + Walkthrough + Diagram + Gotchas + Trace
Deep	"deep dive", "everything", "teach me"	All above + Context Map + Complexity Notes

4. Deliver explanation

Summary (all depths): 1-2 technical sentences -- what it does and why it exists.

Walkthrough (all depths): numbered steps with file:line references, following execution order. Brief: 3-5 steps max.

Diagram (standard+deep): ASCII art reflecting actual code structure. Choose type: data flow, call chain, state machine, sequence. Width under 70 chars. No Mermaid.

Gotchas (standard+deep): specific pitfalls in this code -- edge cases, performance traps, concurrency hazards. Not generic advice.

Trace It (standard+deep): execution trace through a concrete scenario. Show data transformation at each step, highlight decision points.

Context Map (deep only): when to use, when NOT to use, alternatives with tradeoff comparison.

Complexity Notes (deep only): cyclomatic complexity, nesting depth, time/space complexity of hot paths, concurrency assessment.

5. Follow-up

"What about X?" -- extend at same depth.
"Go deeper" -- increase one level, deliver only new sections.
"Trace a different path" -- re-run Trace with different scenario.
"Show me in my code" -- find the concept via Grep/Glob.

Quick Reference

/ai-explain <topic>                  # standard depth (default)
/ai-explain <topic> --depth brief    # summary + walkthrough only
/ai-explain <topic> --depth deep     # all sections including complexity

Common Mistakes

Over-explaining standard concepts -- assume technical competence.
Using Mermaid diagrams -- always ASCII art for terminal readability.
Generic gotchas ("be careful with null") -- must be specific to the code under review.
Using "basically", "simply", "just" -- these minimize real complexity.

Integration

Shared by /ai-guide for teaching mode.
References actual codebase via Grep/Glob, not idealized examples.
Read-only -- never writes code, tests, or documentation.

References

.claude/skills/ai-guide/SKILL.md -- uses explain for teaching interactions.
.claude/skills/ai-verify/SKILL.md -- architecture assessment mode. $ARGUMENTS

---name: ai-explaindescription: Use when explaining code, concepts, patterns, or architecture: engineer-grade technical explanations with 3-tier depth control, ASCII diagrams, and execution traces.--- # Explain Engineer-grade technical explanations of code, concepts, patterns, and architecture. 3-tier depth control scales detail to what the developer needs. Anchored in the actual codebase with `file:line` references, ASCII diagrams, and execution traces. ## When to Use - "How does this work?", "What is this?", "Why does this do X?", "Trace this." - NOT for generating documentation -- use `/ai-write`. - NOT for writing/fixing code -- use `ai-build agent` or `/ai-debug`. ## Process ### 1. Identify subject Classify into: Code, Concept, Pattern, Architecture, Error, or Difference. If ambiguous, ask ONE clarifying question. ### 2. Search codebase Use Grep/Glob to find real instances. Codebase examples with `file:line` references are primary evidence. If not found, use generic example in project's stack and note it. ### 3. Select depth | Depth | Trigger cues | Sections | |-------|-------------|----------| | Brief | "TL;DR", "brief", "short" | Summary + Walkthrough (3-5 steps) | | Standard | General question (DEFAULT) | Summary + Walkthrough + Diagram + Gotchas + Trace | | Deep | "deep dive", "everything", "teach me" | All above + Context Map + Complexity Notes | ### 4. Deliver explanation **Summary** (all depths): 1-2 technical sentences -- what it does and why it exists. **Walkthrough** (all depths): numbered steps with `file:line` references, following execution order. Brief: 3-5 steps max. **Diagram** (standard+deep): ASCII art reflecting actual code structure. Choose type: data flow, call chain, state machine, sequence. Width under 70 chars. No Mermaid. **Gotchas** (standard+deep): specific pitfalls in this code -- edge cases, performance traps, concurrency hazards. Not generic advice. **Trace It** (standard+deep): execution trace through a concrete scenario. Show data transformation at each step, highlight decision points. **Context Map** (deep only): when to use, when NOT to use, alternatives with tradeoff comparison. **Complexity Notes** (deep only): cyclomatic complexity, nesting depth, time/space complexity of hot paths, concurrency assessment. ### 5. Follow-up - "What about X?" -- extend at same depth. - "Go deeper" -- increase one level, deliver only new sections. - "Trace a different path" -- re-run Trace with different scenario. - "Show me in my code" -- find the concept via Grep/Glob. ## Quick Reference ``` /ai-explain <topic> # standard depth (default) /ai-explain <topic> --depth brief # summary + walkthrough only /ai-explain <topic> --depth deep # all sections including complexity ``` ## Common Mistakes - Over-explaining standard concepts -- assume technical competence. - Using Mermaid diagrams -- always ASCII art for terminal readability. - Generic gotchas ("be careful with null") -- must be specific to the code under review. - Using "basically", "simply", "just" -- these minimize real complexity. ## Integration - Shared by `/ai-guide` for teaching mode. - References actual codebase via Grep/Glob, not idealized examples. - Read-only -- never writes code, tests, or documentation. ## References - `.claude/skills/ai-guide/SKILL.md` -- uses explain for teaching interactions. - `.claude/skills/ai-verify/SKILL.md` -- architecture assessment mode. $ARGUMENTS

Frequently asked questions