kaoyan

822电子技术基础考研学习入口。用于湖南大学822专业课的新学、题型训练、错题回收、章节复盘、模电/数电进度推进。默认围绕“章节输入 -> 题型SOP -> 错题归因 -> 进度更新”的考研闭环执行。

822电子技术基础知识结构与章节复盘模块。用于模电/数电章节定位、前置知识判断、高频题型映射、学习顺序建议、章节卡片和进度骨架生成。

This skill handles note generation and updates for 考研数学 (Chinese graduate entrance math exam) preparation. Use it when users want to generate exam-oriented study notes from existing materials, update notes based on feedback, or create structured learning content with LaTeX formatting.

This skill handles note generation and updates for 考研数学 (Chinese graduate entrance math exam) preparation. Use it when users want to generate exam-oriented study notes from existing materials, update notes based on feedback, or create structured learning content with LaTeX formatting.

审计并优化 LLM 提示缓存命中率、输入 token、延迟与调用成本。用于用户要求“优化缓存命中”“降低 token 成本”“审计 LLM 调用”“提示词缓存优化”“优化 AI 调用费用”，或需要为任意 agent profile 与应用代码建立可观测性和固定回归样本时。

Create or retrofit reusable named workflow state machines for multi-step agent projects. Use when a project needs recoverable workflow state files, multiple workflow definitions, phase gates, restart-safe agent workflows, explicit skipped/blocked states, workflow routing rules, or reusable workflow templates across repositories.

自我学习阶段。回顾本次学习会话，记录学习心得和错误到 .learnings/，当文件超阈值时自动压缩去重，更新 RULES.md，促进系统持续改进。触发时机：用户审核通过 evaluate 产出并明确要求记录学习后，Phase 6。

维护 .learnings/ 经验库，把过多或反复出现的学习记录、错误日志、规则失效问题聚类诊断，追溯并修改对应 skill、模板、hook、校验脚本或项目规则；修复并验证后再归档或移除已解决记录；同时检查多个 agent skill 目录之间的共享功能同步。用户提到 learnings 太多、错误反复犯、清理经验库、维护自我学习、压缩错误日志、从错误中修技能、同步多 agent 技能时触发。

维护 .learnings/ 经验库，把过多或反复出现的学习记录、错误日志、规则失效问题聚类诊断，追溯并修改对应 skill、模板、hook、校验脚本或项目规则；修复并验证后再归档或移除已解决记录；同时检查多个 agent skill 目录之间的共享功能同步。用户提到 learnings 太多、错误反复犯、清理经验库、维护自我学习、压缩错误日志、从错误中修技能、同步多 agent 技能时触发。

自我学习阶段。回顾本次会话，记录真实发生的学习点和错误到 .learnings/；当经验库过长时压缩去重并更新 RULES.md；如果发现重复错误或规则失效，转交 maintain-learnings 先修源头。用户明确要求记录学习、复盘、写入 learnings、digest 时触发。

Audit a Git repository for exposed API keys, tokens, passwords, private keys, and other credentials without printing their values. Use when asked to check repository security, scan for leaked secrets, review files before committing or pushing, investigate a credential leak, or check Git history after a possible exposure.

This skill handles review planning and progress tracking for 考研英语 (Chinese graduate entrance English exam) vocabulary learning. Use it when users want to generate spaced repetition schedules based on the SM-2 algorithm, track vocabulary statistics, get daily review lists, or analyze learning progress with phase-based strategies.

This skill provides knowledge point structure templates and module organization for 考研数学 (Chinese graduate entrance math exam). Use it when users want to query the directory structure of high math/linear algebra/probability, get knowledge point relationship graphs, or understand chapter organization.

使用 Python + Matplotlib 生成教科书级别的数学函数图像。当用户想要画函数图、生成图像、可视化数学概念、替换 ASCII 字符图时使用此技能。支持任意函数表达式、多图对比、专业标注。

This skill helps users quickly organize mistakes/errors into subject-specific mistake notebooks. It supports multiple subjects (math, electronics, English), auto-formats mistake entries following the established template with LaTeX support, auto-updates the index table, and appends to existing mistake notebooks without overwriting. Use when user says "整理错题", "记错题", "错题笔记", "把这道题记到错题本", or provides mistake content for recording.

错题本结构优化技能。自动分析错题关联、按知识点分类索引、生成关联网络图。触发词："重构错题本"、"整理错题结构"、"优化错题索引"。

技能自动化重构器 - 自动提取代码到code.md、拆分过长内容、优化技能结构。**自动触发**：每次修改.Codex/skills目录下的SKILL.md或code.md后检测并提示用户确认。手动触发：当用户提到"重构技能"、"拆分技能"、"技能代码分离"、"优化技能结构"时使用此skill。

This skill should be used when the user wants to process Word document templates, extract placeholders, or generate Word documents based on templates.

电路图解析 - 管理822电子技术基础的电路图智能识别、元件参数提取、电路拓扑分析、静态分析+动态分析输出，使用MCP工具实现电路结构识别，强制康华光符号体系

核心协调层 - 管理822电子技术基础的MemOS集成、调度信号处理、统一错误模型、跨学科知识关联（数学↔电子技术）、数学前置检查、考点权重与优先级算法

This skill routes 822 electronics learning requests to specialized sub-skills for 湖南大学822电子技术基础考研 preparation, including circuit diagram analysis, SOP templates for 17 problem types, knowledge point structure, and MemOS integration for persistent tracking.

SOP模板库 - 管理822电子���术基础的17个标准化解题流程（模电8个+数电9个），包含解题步骤引导、答题检查清单、康华光符号体系强制、LaTeX电子符号标准、Mermaid波形图生成

知识点结构 - 管理822电子技术基础的知识点图谱（模电+数电）、前置知识关联、跨章节提示、知识点卡片模板、常见错误模式，复用knowledge_graph_electronics.yaml和knowledge_card_electronics.md

This skill routes English vocabulary learning requests to specialized sub-skills for 考研英语 (Chinese graduate entrance English exam) preparation. It handles vocabulary organization from PDF exports, spaced repetition schedules, quizzes, polysemy (rare word meanings) detection, word lookup, and writing output practice with MemOS integration for persistent tracking.

This skill handles note generation and updates for 考研数学 (Chinese graduate entrance math exam) preparation. Use it when users want to generate exam-oriented study notes from existing materials, update notes based on feedback, or create structured learning content with LaTeX formatting.

This skill routes mathematics learning requests to specialized sub-skills for 考研数学 (Chinese graduate entrance math exam) preparation, including note generation with LaTeX formatting, knowledge point structure templates, and core infrastructure with MemOS integration for persistent mistake tracking and cross-device synchronization.

This skill should be used when the user asks to generate study plans for 考研 (Chinese graduate entrance exam), parse course schedules, create daily/weekly study schedules, or optimize study time allocation. Supports three input modes (minimal/standard/advanced), adapts to individual chronotypes (morning person/night owl), handles task debt from missed plans with circuit breaker protection (>10h triggers recovery mode), enforces Sunday review, respects minimum block duration requirements for different subjects, implements science-based time block splitting based on cognitive science (attention decay, decision fatigue, spacing effect), integrates with MemOS for persistent learning progress tracking, includes context refresh mechanism (auto-prompts profile update after 30 days), mental health intervention (triggers after 3 consecutive tired days), vocabulary review validation (prevents missing newly learned vocabulary), plan upsert logic with tagging for version control, completion record generation based on user