explain

Builds grounded explanations of design documents, plans, issues, code, and technical concepts.

Updated Sep 10, 2026
One-click install
npx skills add https://github.com/sergeiwallace/skills --skill explain-sergeiwallace
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: explain
Source: https://github.com/sergeiwallace/skills/tree/main/skills/explain
Command: npx skills add https://github.com/sergeiwallace/skills --skill explain-sergeiwallace

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Explanations often compress content into summaries or invent rationale that was never recorded. This Skill builds a faithful mental model of a design document, plan, issue, code target, or technical concept by grounding every claim in inspected evidence and labeling provenance. ## Core Features & Use Cases - Evidence-grounded explanation: Inspects the target and minimum surrounding evidence, then marks claims as Observed, Recorded rationale, Inference, or Unknown. - Five input adapters: Tailored procedures for design documents, plans, issues, code, and technical concepts, each with explicit do-not rules. - Calibrated depth: Honors explicit depth, audience, format, and learning-mode signals, defaulting to a technically literate but unfamiliar reader. - Use Case: Ask why a design chose polling over events; the Skill distinguishes recorded rationale from inference, names what is unknown, and cites the exact evidence inspected. ## Quick Start Ask the agent to explain why the release plan verifies before tagging, or to trace a specific function through one timeout scenario.

Frequently Asked Questions about explain

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

FAQPage Schema
How do I get an explanation of a design document or plan?

Invoke /explain with the document path or subject and an optional goal, depth, or audience signal. The Skill inspects the document's decisions, constraints, and alternatives, then explains the causal chain with evidence references.

How do I trace what a function or code path does?

Provide a resolvable file, symbol, or code selection and ask for a trace. The Skill inspects the implementation, callers, tests, and configuration, then walks one concrete execution path rather than inferring behavior from names.

What is the difference between explain and summarize?

Summarize compresses a selected evidence set, while explain closes a mental-model gap by exposing mechanism, rationale, boundaries, and consequences. The Skill explicitly refuses to act as a summarization shortcut.

What happens when the rationale for a decision is not recorded?

The Skill labels the best supported account as Inference or states Unknown instead of inventing author intent. It may suggest targeted Git history or linked research as a deeper route.

Can it explain general technical concepts to a novice?

Yes, it uses a conservative default audience of technically literate but unfamiliar readers, and can add a bounded analogy with an explicit mapping and stopping point. Stable definitions are kept separate from repository-local usage.

What are the limitations when issue data is unavailable?

If live issue fields cannot be accessed, the Skill discloses that limit and never treats a related edge as a blocking dependency. It can explain a supplied snapshot only when labeled with its observed date or revision.