Chess Coach
This skill helps you improve your chess game by monitoring your Chess.com profile and offering advice tailored to your skill level.
Features
- - Profile Tracking: Connect your Chess.com username and track stats across formats (Blitz, Bullet, Rapid).
- Recent Game Analysis: Analyze your most recent games and identify coaching opportunities.
- Skill-Level Calibration: Automatically Adjusts advice based on your current rating.
Workflows
1. Setup
When a user mentions chess or improving their game for the first time:
- 1. Ask for their Chess.com username.
- Use
scripts/chess_api.py <username> stats to fetch their rating and stats. - Store the username and their preferred formats (Blitz, Bullet, Rapid) in
memory/chess_state.json.
2. Monitoring & Analysis
Periodically or upon request:
- 1. Fetch recent games using
scripts/chess_api.py <username> games. - Compare the game list against
memory/chess_state.json to ensure no games are analyzed twice. - Daily Budget: Only analyze one game per 24-hour period unless specifically requested.
- If new games exist, pick the most relevant one (e.g., a loss where the rating gap was small) and analyze it.
- Observation Logging: Instead of just giving a one-off tip, log the core mistake or pattern (e.g., "Missed pin on d-file", "Weakened back rank") to
memory/chess_observations.jsonl. - Consult
references/skill_levels.md for coaching focus based on the user's current rating. - Provide 2-3 actionable points. Keep it punchy and encouraging.
3. Synthesis (The "Grand Lesson")
Once per week, or when memory/chess_observations.jsonl reaches 5+ entries:
- 1. Review the logged observations to find recurring themes.
- Identify the single "highest ROI" area for improvement (e.g., "You lose 40% of games due to early Queen excursions").
- Present this as a dedicated "Coaching Review" session.
Example
"I see you've played 10 Blitz games since we last talked. In your game against 'Grandmaster123', you missed a simple fork on the f2 square! At your current rating (850), focusing on basic tactics like these will help you reach 1000 fast."
Resources
- -
scripts/chess_api.py: Interacts with the Chess.com public API. references/skill_levels.md: Provides guidance on coaching priorities by rating bracket.memory/chess_state.json: Store the user's username, preferred formats, and the ID of the last analyzed game.
国际象棋教练
该技能通过监控您的Chess.com个人资料，并根据您的技能水平提供定制建议，帮助您提升棋艺。
功能特点
- - 个人资料追踪：关联您的Chess.com用户名，追踪各模式（快棋、子弹棋、速棋）数据。
- 近期对局分析：分析您最近的对局，发现可改进之处。
- 水平校准：根据您当前等级分自动调整建议内容。
工作流程
1. 初始设置
当用户首次提及国际象棋或提升棋艺时：
- 1. 询问其Chess.com用户名。
- 使用scripts/chessapi.py <用户名> stats获取等级分和数据。
- 将用户名和偏好模式（快棋、子弹棋、速棋）存储至memory/chessstate.json。
2. 监控与分析
定期或按需执行：
- 1. 使用scripts/chessapi.py <用户名> games获取近期对局。
- 将对局列表与memory/chessstate.json比对，避免重复分析。
- 每日限额：除非特别要求，每24小时仅分析一局对局。
- 如有新对局，选取最具参考价值的一局（如等级分差距较小的失利局）进行分析。
- 观察记录：不提供一次性建议，而是将核心失误或模式（如漏看d线牵制、削弱底线）记录至memory/chessobservations.jsonl。
- 参考references/skilllevels.md，根据用户当前等级分确定指导重点。
- 提供2-3个可操作要点，保持简洁且具有鼓励性。
3. 综合总结（大师课）
每周一次，或当memory/chess_observations.jsonl达到5条以上记录时：
- 1. 回顾记录的观察结果，找出重复出现的主题。
- 确定最具投资回报率的改进方向（如您40%的失利源于过早出动皇后）。
- 以专门的教练复盘形式呈现。
示例
自上次交流以来，您已下了10盘快棋。在与Grandmaster123的对局中，您漏看了f2格上简单的捉双！以您当前850的等级分，专注于这类基础战术将助您快速突破1000分。
资源
- - scripts/chessapi.py：与Chess.com公共API交互。
- references/skilllevels.md：按等级分段提供指导重点建议。
- memory/chess_state.json：存储用户名、偏好模式及最后分析对局ID。
