aps-explanation-editorial

A good editorial restates the problem formally, connects constraints to solution requirements, shows the brute-force baseline and why it fails, walks the insight path to the better approach, and includes correctness reasoning, complexity analysis, implementation notes, and common pitfalls.

Hints sit on a 7-level ladder from meta nudges to key insights and are given during solving to preserve discovery, while editorials are complete explanations usually unlocked post-solve or opt-in. They carry different spoiler risks and learning impacts, so products should separate them.

Reveal policy should be a function of mode, solver level, and struggle time. Learning mode favors facilitative, delayed hints with struggle gates before higher levels unlock, while assessment mode strictly withholds explanations to certify results.

Viewing an editorial before any attempt signals high assistance and potential passive learning, while viewing after independent AC supports review. The strongest signal is when the user reconstructs or explains the solution in their own words after reading.

AI explanations must be gated on having concrete diagnostic evidence rather than generic feedback, since confident but wrong or generic explanations damage trust and learning. This Skill coordinates with an AI Assistance & Integrity Expert for guardrails.

It does not solve algorithm problems, write competitive-programming code, or debug user submissions. It produces product requirements, domain distinctions, and metrics for planners and engineers rather than implementation details like schemas or APIs.