Back to Details

token-coach

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Token Coach: Plan Token-Efficient Before You Build

Token Coach：在构建前规划Token高效方案

Interactive coaching for Claude Code architecture decisions. Analyzes your setup, identifies patterns (good and bad), and gives personalized advice with real numbers.
Use when: Building something new, existing setup feels slow, designing multi-agent systems, or want a quick health check.
针对Claude Code架构决策的交互式指导工具。分析你的配置，识别优劣模式，并结合真实数据给出个性化建议。
适用场景：构建新项目、现有配置运行缓慢、设计多Agent系统，或需要快速进行健康检查时。

Phase 0: Initialize

阶段0：初始化

  1. Resolve measure.py path (same as token-optimizer):
bash
MEASURE_PY=""
if [ -f "$HOME/.claude/skills/token-optimizer/scripts/measure.py" ]; then
  MEASURE_PY="$HOME/.claude/skills/token-optimizer/scripts/measure.py"
else
  MEASURE_PY="$(find "$HOME/.claude/plugins/cache" -path "*/token-optimizer/scripts/measure.py" 2>/dev/null | head -1)"
fi
[ -z "$MEASURE_PY" ] || [ ! -f "$MEASURE_PY" ] && { echo "[Error] measure.py not found. Is Token Optimizer installed?"; exit 1; }
  1. Collect coaching data:
bash
python3 $MEASURE_PY coach --json
Parse the JSON output. This gives you: snapshot (current measurements), detected patterns, coaching questions, and focus suggestions.
  1. Check context quality (v2.0):
bash
python3 $MEASURE_PY quality current --json 2>/dev/null
If available, parse the quality score and issues. This enriches coaching with session-level insights (not just setup overhead). If the command fails (pre-v2.0 install), skip gracefully.
  1. 解析measure.py路径（与token-optimizer路径一致）：
bash
MEASURE_PY=""
if [ -f "$HOME/.claude/skills/token-optimizer/scripts/measure.py" ]; then
  MEASURE_PY="$HOME/.claude/skills/token-optimizer/scripts/measure.py"
else
  MEASURE_PY="$(find "$HOME/.claude/plugins/cache" -path "*/token-optimizer/scripts/measure.py" 2>/dev/null | head -1)"
fi
[ -z "$MEASURE_PY" ] || [ ! -f "$MEASURE_PY" ] && { echo "[Error] measure.py not found. Is Token Optimizer installed?"; exit 1; }
  1. 收集指导数据：
bash
python3 $MEASURE_PY coach --json
解析JSON输出。你将获得：快照（当前测量数据）、检测到的模式、指导问题及重点建议。
  1. 检查上下文质量（v2.0版本）：
bash
python3 $MEASURE_PY quality current --json 2>/dev/null
如果可用，解析质量分数和问题。这会为指导增加会话级别的洞察（不仅是配置开销）。如果命令执行失败（安装版本低于v2.0），请优雅跳过。

Phase 1: Intake

阶段1：需求收集

Ask ONE question:
What's your goal today? a) Building something new, want it token-efficient from the start b) Existing project feels sluggish / context fills too fast c) Designing a multi-agent system, want architecture advice d) Quick health check with actionable tips
Wait for the answer. Don't dump info before they choose.
提出一个问题：
你今天的目标是什么？ a) 构建新项目，希望从一开始就实现Token高效 b) 现有项目运行缓慢 / 上下文填充过快 c) 设计多Agent系统，需要架构建议 d) 快速健康检查并获取可操作建议
等待用户回答。在用户选择前不要输出大量信息。

Phase 2: Load Context (based on intake)

阶段2：加载上下文（基于需求收集结果）

Resolve the token-coach skill directory:
bash
COACH_DIR=""
if [ -d "$HOME/.claude/skills/token-coach" ]; then
  COACH_DIR="$HOME/.claude/skills/token-coach"
elif [ -d "$HOME/.claude/skills/token-optimizer/../token-coach" ]; then
  COACH_DIR="$HOME/.claude/skills/token-optimizer/../token-coach"
else
  COACH_DIR="$(find "$HOME/.claude/plugins/cache" -path "*/token-coach" -type d 2>/dev/null | head -1)"
fi
Load references based on intake choice:
  • Option a or b: Read 
    $COACH_DIR/references/coach-patterns.md
    + 
    $COACH_DIR/references/quick-reference.md
  • Option c: Read 
    $COACH_DIR/references/agentic-systems.md
    + 
    $COACH_DIR/references/quick-reference.md
  • Option d: Read 
    $COACH_DIR/references/quick-reference.md
    only (fast path)
Read the matching example from 
$COACH_DIR/examples/
as a few-shot template:
  • Option a: 
    coaching-session-new-project.md
  • Option b: 
    coaching-session-heavy-setup.md
  • Option c: 
    coaching-session-agentic.md
  • Option d: Skip example (keep it fast)
Read 
$COACH_DIR/references/coaching-scripts.md
for conversation structure.
解析token-coach技能目录：
bash
COACH_DIR=""
if [ -d "$HOME/.claude/skills/token-coach" ]; then
  COACH_DIR="$HOME/.claude/skills/token-coach"
elif [ -d "$HOME/.claude/skills/token-optimizer/../token-coach" ]; then
  COACH_DIR="$HOME/.claude/skills/token-optimizer/../token-coach"
else
  COACH_DIR="$(find "$HOME/.claude/plugins/cache" -path "*/token-coach" -type d 2>/dev/null | head -1)"
fi
根据用户选择加载参考资料：
  • 选项a或b：读取
    $COACH_DIR/references/coach-patterns.md
    + 
    $COACH_DIR/references/quick-reference.md
  • 选项c：读取
    $COACH_DIR/references/agentic-systems.md
    + 
    $COACH_DIR/references/quick-reference.md
  • 选项d：仅读取
    $COACH_DIR/references/quick-reference.md
    （快速路径）
从
$COACH_DIR/examples/
中读取匹配的示例作为少样本模板：
  • 选项a：
    coaching-session-new-project.md
  • 选项b：
    coaching-session-heavy-setup.md
  • 选项c：
    coaching-session-agentic.md
  • 选项d：跳过示例（保持快速）
读取
$COACH_DIR/references/coaching-scripts.md
以获取对话结构。

Phase 3: Coach (conversation, not report)

阶段3：指导（对话形式，而非报告）

This is a CONVERSATION. Not a wall of text.
  1. Lead with the 1-2 most impactful findings from the coaching data
  2. If quality data is available and score < 70, lead with that instead: "Your current session quality is [X]/100. [Top issue] is eating [Y tokens]."
  3. Reference their actual numbers ("You have 47 skills costing ~4,700 tokens at startup")
  4. Ask a follow-up question. Don't dump everything at once.
  5. For agentic systems (option c): walk through their architecture step by step
  6. Use the coaching scripts for structure, but keep it natural
Tone: Knowledgeable friend, not corporate consultant. Be direct about what matters and why. Use real numbers from their data.
Anti-patterns to call out: Reference the anti-patterns from coach-patterns.md. Name them ("You've got the 50-Skill Trap going on").
Continue the conversation for 2-4 exchanges. Let the user ask questions. Adjust advice based on what they tell you about their workflow.
这是一场对话，不是长篇大论。
  1. 先从指导数据中提出1-2个最具影响力的发现
  2. 如果质量数据可用且分数低于70，则优先说明："你当前的会话质量为[X]/100。[首要问题]消耗了[Y]个Token。"
  3. 引用真实数据（"你有47个技能，启动时约消耗4700个Token"）
  4. 提出后续问题。不要一次性输出所有内容。
  5. 对于Agent系统（选项c）：逐步梳理其架构
  6. 参考指导脚本的结构，但保持对话自然
语气：像知识渊博的朋友，而非企业顾问。直接说明关键问题及其原因。使用数据中的真实数字。
需要指出的反模式：参考coach-patterns.md中的反模式，直接点名（"你陷入了50技能陷阱"）。
持续对话2-4轮。允许用户提问。根据用户告知的工作流程调整建议。

Phase 4: Action Plan

阶段4：行动计划

After the conversation, generate a prioritized action plan:
  1. Summarize 3-5 concrete actions, ordered by impact
  2. Include estimated token savings for each action (use the numbers from quick-reference.md)
  3. If quality score < 70: include "Set up Smart Compaction" as a recommended action (
    python3 $MEASURE_PY setup-smart-compact
    )
  4. If quality score < 50: recommend immediate 
    /compact
    or 
    /clear
    before continuing
  5. Flag which actions are quick wins vs deeper changes
  6. Offer to run 
    /token-optimizer
    for the full audit + implementation if they want to go beyond coaching
Format: Keep it scannable. Numbered list with bold action names, one-line description, estimated savings.
对话结束后，生成优先级明确的行动计划：
  1. 总结3-5项具体行动，按影响力排序
  2. 包含每项行动的预估Token节省量（参考quick-reference.md中的数据）
  3. 如果质量分数低于70：将"设置智能压缩"列为推荐行动（
    python3 $MEASURE_PY setup-smart-compact
    ）
  4. 如果质量分数低于50：建议立即执行
    /compact
    或
    /clear
    后再继续
  5. 标记哪些行动是快速见效的，哪些是深度改进
  6. 如果用户希望超越指导范围，提供运行
    /token-optimizer
    进行完整审计和实施的选项
格式：保持易读性。使用编号列表，包含加粗的行动名称、一行描述和预估节省量。

Phase 5: Dashboard (optional)

阶段5：仪表板（可选）

If measure.py generated a coach dashboard tab, mention it: "Your Token Health Score and pattern analysis are in the dashboard. Run 
python3 $MEASURE_PY dashboard
to see it."
如果measure.py生成了指导仪表板标签页，请告知用户： "你的Token健康分数和模式分析已在仪表板中。运行
python3 $MEASURE_PY dashboard
即可查看。"