By name: physics-adjust description: "依据最近学习表现动态调整物理学习计划（难度、节奏、复习占比）。" homepage: https://your-docs/skills/physics-adjust metadata: openclaw: emoji: "⚙️" output_schema: "schemas/adjust.schema.json" policy: "policies/education-rules.json"
Physics Adjust Skill
When to Use
- 每日学习结束后自动微调次日计划
- 连续低正确率/高耗时时需要降阶回补
- 连续高正确率时需要提速或提难
Required Inputs
- student_id
- recent_window (最近7-14天表现)
- current_plan_snapshot
- mastery_map
Core Score
S = 0.5accuracy + 0.2(1-time_ratio) + 0.2retention + 0.1consistency
Adjustment Strategy
- S < 0.45: 降难度，回退到先修点，增加讲解
- 0.45 <= S < 0.70: 保持难度，增加同类巩固练习
- S >= 0.70: 小幅升难并推进新知识点
Output Contract
- 必须返回符合 schemas/adjust.schema.json 的 JSON
- 必须包含 before_after_diff、reasons、risk_flags
- 难度日变化不能超过策略阈值
Safety & Quality Rules
- 连续失败时必须触发“减负+回补”
- 避免计划频繁抖动（设置冷却期）
- 调整必须可解释并可人工覆盖