← ClaudeAtlas

explainlisted

Build a grounded mental model of a design or plan document, issue, code target, or technical concept. Use for explanation, refresher, rationale, comparison, trace, consequence, or learning-mode requests where faithful relationships matter more than compression.
sergeiwallace/skills · ★ 0 · AI & Automation · score 73
Install: claude install-skill sergeiwallace/skills
# explain Give the user enough structure, mechanism, rationale, boundaries, and consequences to reason about the subject. Do not substitute a summary: `/summarize` compresses a selected evidence set, while this skill closes a mental-model gap. Do not invoke `/summarize` as an implementation shortcut or create a shared runtime between the skills. **Usage:** `/explain <subject and optional goal, depth, audience, format, or learning-mode signal>` ## References Read [input-adapters.md](references/input-adapters.md) after resolving the primary adapter and before grounding the answer. Read [examples-and-evals.md](references/examples-and-evals.md) when an example helps select response depth, when evaluating this skill, or when a failure path is in doubt. Keep both references one level from this file; do not load unrelated adapters or examples by default. ## Procedure 1. **Resolve.** Identify the subject, the user's goal, one primary input adapter, and every explicit depth, familiarity, format, audience, or learning-mode signal. Let an explicit type or direct target win; otherwise resolve a readable path, issue ID, selection, or symbol. Treat a bare term as a concept unless repository evidence establishes a project-specific use. 2. **Calibrate.** Explicit depth, familiarity, format, and learning-mode signals win over defaults. Otherwise assume a reader who is technically literate but unfamiliar with this subject. Do not ask an intake question when the target is r