skills

大學生活教練 for Taiwanese university students, built on Cal Newport's "How To Win At College" (75 proven strategies, localized to Taiwan's academic system). Use whenever a college student asks about studying, exams, papers, course selection, time management, campus life, motivation, or feeling lost — even if they never say the word "coach". Trigger phrases: 期中考/期末考怎麼準備、考試範圍好大、 報告寫不完、報告拖到最後、論文不知道怎麼開始、選課/加退選怎麼選、要不要停修、想拚書卷獎、 時間管理好爛、每天都好累、期中低潮、一直熬夜、大學好迷惘、暑假要幹嘛、實習怎麼找, "how should I study for finals", "I keep procrastinating on my paper". Responds as a coach: asks about the student's actual situation first, then gives 1–3 concrete next actions with a plan. NOT for K-12 / 高中升學, graduate thesis advising, requests to ghost-write the assignment or report itself（幫我寫報告／代寫 作業 — the coach guides the student's own writing, never produces it for them）, or mental-health crises (those get referred to campus counseling).

Run a local multi-model code review through the user's OpenCode setup BEFORE a pull request is opened. Use this right after finishing an implementation or bug fix in Claude Code, while the changes are still local — uncommitted, or committed on a branch but not yet turned into a PR — and the user wants an independent review pass before pushing. Trigger whenever the user says things like "review my changes", "review before I open the PR", "run the opencode review", "multi-model review", "review this diff", "second opinion on this", or asks for a pre-PR review of work that was just completed, even if they don't name OpenCode explicitly. Returns a consolidated review report on stdout for Claude Code to act on. Do NOT use this for reviewing an already-open GitHub PR — that path is handled interactively in the OpenCode TUI and uses gh.

Use when a bug is fixed, a deploy completes, tests go green, the user encounters a classic pitfall, the user explicitly asks for a joke, or the conversation calls for levity. Also fits when discussing math proofs, physics problems, data analysis, or any STEM topic where a well-placed joke would be welcome. Covers programming, math, physics, chemistry, biology, economics, philosophy, linguistics, engineering, astronomy, and statistics. Trigger phrases: 說個笑話、來個冷笑話、講笑話、tell me a joke, need a laugh. Do NOT use mid-explanation or when the user is frustrated and wants a solution.

Use this skill whenever the user wants test code written, added, audited, or judged "good enough" — not merely run. Trigger when they: ask you to write or improve tests for a module (parser, retry/network logic, migration, auth/token, billing, state machine, library/SDK); worry that edge cases or "garbage input" keep slipping through release after release; ask what testing a component needs before shipping, open-sourcing, or tagging a version; want code made robust, reliable, or production-ready; report flaky, missing, or weak tests; or raise data corruption, crash/rollback/atomicity, concurrency, or fault-handling concerns. Also apply proactively right after non-trivial logic is implemented, to cover failure paths and boundaries even if "test" was never said. Brings SQLite-grade methods (fuzzing, property/differential/metamorphic testing, fault injection, mutation testing, coverage) but right-sizes rigor to the component's blast radius. Not for just running an existing suite or general debugging.