本文へ移動
cccskills
無料GitHub で公開

revision-mode

Use when a learner needs a structured review session prioritizing high-value topics by recency, error history, and upcoming deadlines.

インストール方法を見る

含まれるファイル(3)

  • SKILL.md3.6 KB
  • examples.md1.7 KB
  • README.md764 B

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Purpose

Run a structured, prioritized revision session targeting topics most likely to decay, most frequently missed, or most important for an upcoming deadline. Not random review — weighted prioritization to maximize learning per minute.

Activation

  • Upcoming deadline (interview, exam, demo) within 1–5 sessions. Learner requests review after a learning period. weak-area-tracker shows accumulated weak areas. 3+ sessions since last review.
  • Skip if: learner needs new material first. Goal is project implementation → build-with-me. Only one topic to review → use check-understanding/challenge-generator directly.
  • Routing: pull priorities from weak-area-tracker. Use check-understanding and challenge-generator as execution vehicles. After 3 consecutive clean revision passes on a topic: mark revision-complete and de-prioritize.

Inputs

  • Topics covered (from learning-memory), active weak areas, upcoming deadline + type, session time budget, self-assessed confidence per topic (optional).

Priority Scoring (1–10)

  • Weak area severity: 0–3 (from tracker). Recency (days since last review): 0–3. Deadline relevance: 0–2. Low self-reported confidence: 0–2.
  • Score 7+ = must-revise. Score 4–6 = should-revise. Below 4 = defer.

Workflow

  1. Scope — Gather topics + weak areas. Ask: "Deadline? Least confident topics?" Apply priority scoring.
  2. Plan — Present focused list: must-revise first, then should-revise if time permits. Cap at 4–5 topics max. State approach per topic: concept check / challenge / flashcard drill.
  3. Recap — For each topic: ask learner to summarize in 2–3 sentences. Don't re-explain unless summary reveals a gap. 2–3 min cap per recap.
  4. Test — 1–2 exercises per topic via check-understanding or challenge-generator. Mix modes. Run 3–5 due flashcards per topic if available.
  5. Check Exit — Clean pass = correct response with correct reasoning on first attempt. After 3 clean passes across sessions: exit rotation. Below 50% pass rate this session: escalate to misconception-detector.
  6. Close — Summarize: which topics are revision-complete, which need another round. Update weak-area-tracker. Recommend next session if deadline still approaching.

Rules

  • DO: revision is active — learner produces answers, not just reads.
  • DO: include weak areas in every revision session until resolved.
  • DO: state clean-pass criteria before the session, not after.
  • DON'T: re-lecture topics the learner can summarize correctly — test immediately.
  • DON'T: pack more than 5 topics per session.
  • DON'T: skip the hardest topics in favor of comfortable ones — must-revise (7+) first.
  • DON'T: let same topics appear indefinitely — enforce exit criteria.

Output

Responses should contain: deadline + time budget, prioritized topic list (must-revise with scores and approach, should-revise if time permits), per-topic recall + test + outcome, and session summary (clean passes, still-needs-revision, exiting rotation, next session recommendation). Format naturally.

Checklist

  • Priority scoring applied before session plan.
  • Must-revise topics addressed first.
  • Every topic has at least one active retrieval exercise.
  • Weak-area-tracker updated at session close.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Use when critiquing a system design across correctness, scalability, reliability, maintainability, and observability using probe-first questioning.

日本語の概要は準備中です。原文の説明を表示しています。

yugash007/edu-agent-skills92026年5月19日 更新

Use when guiding a learner through real project implementation while preserving their ownership and teaching through the build process.

日本語の概要は準備中です。原文の説明を表示しています。

yugash007/edu-agent-skills92026年5月19日 更新

Use when generating personalized practice challenges calibrated to the learner's weak areas, level, and project context.

日本語の概要は準備中です。原文の説明を表示しています。

yugash007/edu-agent-skills92026年5月19日 更新

Use when verifying whether the learner truly understands a concept through reasoning and application, not recall trivia.

日本語の概要は準備中です。原文の説明を表示しています。

yugash007/edu-agent-skills92026年5月19日 更新

Use when coaching hypothesis-driven debugging that requires the learner to gather evidence and reason about root causes before applying fixes.

日本語の概要は準備中です。原文の説明を表示しています。

yugash007/edu-agent-skills92026年5月19日 更新

deep-dive

無料

Use when a learner has surface understanding and needs to explore mechanism, tradeoffs, edge cases, and production implications at depth.

日本語の概要は準備中です。原文の説明を表示しています。

yugash007/edu-agent-skills92026年5月19日 更新

yugash007 のスキルをすべて見る

このスキルの問題を報告する