question-mediumlisted

## When to load this skill The question needs genuine reasoning, file or version verification, or stack-specific knowledge, but does not involve novel architecture, multi-layer debugging, or long correctness chains. This is the default for most substantive questions. ## When not to load this skill Escalate to `/question-hard` when the reasoning chain is long, the tradeoffs are architectural, or a wrong answer is expensive. Drop to `/question-easy` for lookups and recall with a direct path to the answer. ## Procedure 1. Read the question precisely. If $ARGUMENTS is ambiguous on a point that changes the answer, ask one focused clarifying question before proceeding. 2. Verify before answering when the question warrants it. If it concerns a specific file, project, or repo, read the relevant files rather than assuming. If it concerns the current version, features, or API of a fast moving tool, check the lockfile or authoritative source. 3. If the stack matters: the `<repo-context>` block carries it when this runs inline; forked (no hook injection reaches the subagent), so absent that block, identify the stack from the project's own config directly. Load the matching patterns skill if the question is in that area. 4. Answer with the reasoning that supports it. Name a tradeoff or edge case when one is material; do not invent complexity that is not there. ## Output Answer in chat. Not a deliverable command; no scratch artifact unless the user asks. - Direct answer first, then