explain

Answers codebase questions from architecture docs and the knowledge graph without scanning source.

1|Updated Jun 27, 2026
One-click install
npx skills add https://github.com/vickysrawat/AI-Assisted-development --skill explain-vickysrawat
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: explain
Source: https://github.com/vickysrawat/AI-Assisted-development/tree/main/skills/explain
Command: npx skills add https://github.com/vickysrawat/AI-Assisted-development --skill explain-vickysrawat

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Developers waste time grepping through large codebases to answer structural questions like "how does authentication flow through the API" or "where is this filter applied". This Skill answers those questions instantly from pre-built architecture documentation and the codebase knowledge graph, reading at most one source file — and only with explicit consent. ## Core Features & Use Cases - Docs-first Q&A: Loads .claude/architecture/ docs and the .claude/graph/ knowledge graph to answer structural questions without touching source code. - Consent-gated source reading: When docs are insufficient, it identifies the single most relevant file and asks permission before reading exactly one file (Consent Category B). - Dependency-aware answers: Recognizes modules from locally-cloned dependency repos (additionalDirectories) in the multi-root knowledge graph and points to the correct dependency file. - Use Case: A new team member asks "/explain How does Azure AD auth flow through the API layer?" and receives a precise answer naming the relevant module, entry point, and key files — no source scanning required. ## Quick Start Ask a structural question about the codebase, for example: /explain "Where are matter filters applied?"

Frequently Asked Questions about explain

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I ask questions about a codebase without reading source code?

Run /explain followed by your question, such as "/explain How does error handling work?". The Skill answers from architecture docs and the knowledge graph, only reading a source file if you explicitly approve it.

What is a codebase knowledge graph used for in code Q&A?

A knowledge graph maps modules, entry points, and key files so structural questions can be answered without scanning source. The /explain Skill reads `.claude/graph/graph-index.md` and module detail files to locate exactly where functionality lives.

Does /explain work if architecture docs are missing?

No. If neither `.claude/architecture/` docs nor the knowledge graph exist, the Skill stops and instructs you to run /setup-init first to generate them, since it relies on these artifacts to answer without scanning source.

Can /explain read files in dependency repositories?

Yes. When the multi-root knowledge graph indicates the answer lives in a locally-cloned dependency from `additionalDirectories`, the Skill names that dependency module and may read one file there — still gated by your explicit confirmation.

What are the limitations of answering from architecture docs?

Highly specific implementation questions may exceed what architecture docs contain. In those low-confidence cases the Skill names the 1–2 most relevant files and asks before reading, rather than guessing or scanning broadly.