token-reporter

Token Reporter — Daily Token Consumption + Output Reporting

Author: Daniel Li
Copyright © Daniel Li. All rights reserved.

Feature Overview

Automatically tally the local OpenClaw instance's token consumption and Agent work output every day, format it, and write it to Feishu Bitable.

Instance Mapping

Instance	CEO Agent	Reporter
aa (local)	Xiao a (main)	Daniel
Peter Mini	Xiao Tu (xiaotu)	Peter
Xiao m	Xiao m bot	Xiao m

Execution Flow

Step 1: Scan JSONL Logs

Scan incremental data for the day in ~/.openclaw/agents/*/sessions/*.jsonl .

Each assistant message carries a usage object:

{
  "usage": {
    "input": 27813,
    "output": 246,
    "cacheRead": 6720,
    "cacheWrite": 0,
    "totalTokens": 34779,
    "cost": {
      "input": 0, "output": 0,
      "cacheRead": 0, "cacheWrite": 0,
      "total": 0
    }
  }
}

Step 2: Aggregate by Agent + Model

Focus on tracking 4 models (other models are grouped under "Other"):

Model	Matching Rule
opus4.6	`claude-opus-4-6` (xingsuancode, xingjiabiapi)
glm-5	`zai/glm-5`
minimax-M2.5	`minimax/MiniMax-M2.5`
gemini-3-pro	`gemini-3-pro` (xingjiabiapi)

Step 3: Format Token Details

opus4.6: input 320K / output 18K / cache 82% / $4.2
glm-5: input 1.2M / output 85K / cache 78% / $0.3
minimax-M2.5: input 200K / output 12K / cache 0% / $0.1
gemini: — (当日未使用)

cache% = cacheRead / (input + cacheRead + cacheWrite) × 100

Step 4: Collect Output

Summarize each agent's work summary for the day, format:

code: KGKB工厂30/30完成, 5平台发布脚本 | content: 小红书×2, 选题推荐 | ops: cron改造, 服务器巡检

Output collection methods:

Read the day's memory file ~/clawd/memory/$(date +%Y-%m-%d).md
Parse ## Agent名 sections or keyword matching
If there is no memory file, infer from a summary of the last user/assistant message in the JSONL

Step 5: Write to Feishu Bitable

Use the lark-mcp tool to write to Bitable:

Field	Type	Filled By	Description
Employee Name	Text	config	Hardcoded in config.json `person`
TokenDetails	Text	Script statistics	4-model formatted text
Output	Text	Script statistics	Each agent's work summary for the day
Personal Summary	Text	Filled by person	Manually filled by employee (leave blank)
Rating	Single select	Selected by management	Excellent/Good/Fail (leave blank)
Time	Date	Script automatic	Current date

Feishu API

Reference skill: ~/clawd/skills/feishu-automation/SKILL.md
Tool prefix: mcp__lark-mcp_*
Add record: mcp__lark-mcp_createRecord
Query record: mcp__lark-mcp_listRecords

Script Usage

# 统计当日 Token（只看数据，不上报）
python scripts/report.py --scan-only

# 统计 + 上报飞书
python scripts/report.py --report

# 指定日期
python scripts/report.py --date 2026-03-17 --scan-only

# 使用指定配置
python scripts/report.py --config /path/to/config.json --report

Cron Configuration Suggestions

{
  "name": "📊 每日 Token 报告 + 飞书上报",
  "schedule": { "kind": "cron", "expr": "0 23 * * *", "tz": "Asia/Shanghai" },
  "payload": {
    "kind": "agentTurn",
    "message": "执行每日 Token 报告任务：\n1. 运行 python3 ~/clawd/skills/token-reporter/scripts/report.py --scan-only\n2. 查看输出，确认数据合理\n3. 运行 python3 ~/clawd/skills/token-reporter/scripts/report.py --report\n4. 确认飞书写入成功后，用 message(action=send, channel=telegram, target=-1003890797239) 简要汇报\n5. 如果失败，报告错误原因"
  }
}

Configuration

During installation, copy config.example.json → config.json and fill in:

{
  "person": "Daniel",
  "instance": "aa",
  "lark_app_id": "cli_a83467f9ecba5013",
  "bitable_app_token": "your_bitable_token",
  "bitable_table_id": "your_table_id"
}

File Structure

token-reporter/
├── SKILL.md
├── config.example.json    # 配置模板（安装时复制为 config.json）
├── scripts/
│   └── report.py          # 核心统计脚本
└── references/
    └── jsonl-structure.md # JSONL 数据结构参考

Error Handling

Error	Handling
Corrupted JSONL file	Skip that line, log a warning
No data for the day	Report "no data for the day"
Feishu API failure	Retry 3 times; after failure, cache the result locally
config.json does not exist	Prompt to copy config.example.json
Memory file does not exist	Output marked as "no record"