Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
bytedance avatar

Byted Openclaw Diag

  • 30 installs
  • 411 repo stars
  • Updated August 4, 2026
  • bytedance/agentkit-samples

byted-openclaw-diag is a skill that analyzes OpenClaw agent performance, reporting inference latency, token usage, tool-call statistics, run timelines, and errors.

About

byted-openclaw-diag is a diagnostics and performance-analysis tool for OpenClaw agents. It reports model call counts, average inference latency, token throughput, tool-call statistics, per-run timelines, and recent errors. A developer runs it to check an agent's daily run status, find performance bottlenecks, or investigate errors, with a /diag command and modes for summary, full report, per-agent filtering, and live tracking.

  • Analyzes OpenClaw agent inference latency, token usage, and tool-call statistics
  • Modes for batch analysis, live tracking (-f), summary stats (-s), and advanced diagnostics
  • Reads debug logs or session files, degrading gracefully when debug logs are absent

Byted Openclaw Diag by the numbers

  • 30 all-time installs (skills.sh)
  • Ranked #359 of 596 Debugging skills by installs in the Skillselion catalog
  • Data as of Aug 5, 2026 (Skillselion catalog sync)
At a glance

byted-openclaw-diag capabilities & compatibility

Free; runs locally against OpenClaw log and session files with python3 and bash.

Capabilities
performance analysis · agent diagnostics · error analysis
Use cases
debugging
Runs
Runs locally
Pricing
Free
From the docs

What byted-openclaw-diag says it does

无 debug 日志时自动降级为 session 模式,核心指标(推理耗时、Token、工具统计)仍然准确。
SKILL.md
脚本依赖 `python3`(3.7+,使用 `datetime.fromisoformat`)和 `bash`
SKILL.md
- 模型调用次数、平均推理延迟、Token 吞吐量
SKILL.md
npx skills add https://github.com/bytedance/agentkit-samples --skill byted-openclaw-diag

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Installs30
repo stars411
Last updatedAugust 4, 2026
Repositorybytedance/agentkit-samples

What it does

Diagnose OpenClaw agent performance by reporting inference latency, token usage, tool-call stats, and recent errors.

Who is it for?

Investigating OpenClaw agent latency, token consumption, tool-call stats, and errors.

Skip if: Diagnosing non-OpenClaw runtimes, since it parses OpenClaw debug logs and session files.

When should I use this skill?

The user asks about OpenClaw run status, performance bottlenecks, inference latency, token consumption, or errors.

What you get

Produces a summary or full diagnostic report of agent runs, latency, throughput, and recent errors.

  • A summary of model calls, inference latency, and token throughput
  • Per-run timelines and a recent error list

By the numbers

  • Lists the most recent 20 errors in reverse chronological order
  • 2 data sources (debug log, session file)

Files

SKILL.mdMarkdownGitHub ↗

OpenClaw 诊断工具

指令模式

当用户发送 /diag 指令时,直接执行脚本,不做额外解释:

用户输入执行命令说明
/diag-s今日摘要(默认)
/diag full(无-s)完整报告(含 Run 详情 + 错误列表)
/diag full -l 3-l 3最近 3 个 Run 完整详情
/diag -a waicode-s -a waicode指定 agent 摘要
/diag -a main full-a main指定 agent 完整报告
/diag 2026-03-19-s 2026-03-19指定日期摘要
/diag errors(无-s)执行完整报告,只提取错误部分汇总

规则: 1. 无参数时默认 -s(摘要模式,最简洁) 2. full 关键词 → 去掉 -s,输出含 Run 详情 3. errors 关键词 → 执行完整报告,只摘出错误列表 4. -a-l、日期参数直接透传给脚本 5. 去除 ANSI 颜色码:管道 | sed 's/\x1b\[[0-9;]*m//g' 6. 不支持 -f(实时跟踪),该模式需在 SSH 终端运行 7. 直出模式:脚本输出直接用 message 工具原样发送给用户,不经过模型总结。 具体做法:

  • 执行脚本,将 stdout 存入变量
  • message(action="send", message=output) 发送原始输出
  • 然后回复 NO_REPLY(避免重复发送)
  • 如果输出超过 4000 字符,按 4000 字符分段发送(Telegram 消息长度限制)
  • 每段用 ``` 代码块包裹,保持等宽字体排版

自然语言模式

当用户用自然语言询问(如"运行状态怎么样"、"waicode今天干了啥")时, 自行选择合适参数执行脚本,并用中文汇总关键信息。

快速使用

# 诊断今天的数据
bash scripts/byted-openclaw-diag.sh

# 诊断指定日期
bash scripts/byted-openclaw-diag.sh 2026-03-19

# 只看摘要
bash scripts/byted-openclaw-diag.sh -s

# 实时跟踪(类似 tail -f)
bash scripts/byted-openclaw-diag.sh -f

# 高级实时跟踪(自动开启 debug 日志,退出时恢复)
bash scripts/byted-openclaw-diag.sh -f --advanced

# 只看指定 agent
bash scripts/byted-openclaw-diag.sh -a waicode

# 最近 5 个 Run
bash scripts/byted-openclaw-diag.sh -l 5

模式说明

模式参数说明
摘要统计-s(默认)KPI 概览,最简洁
完整报告-s含 Run 详情 + 时间线 + 错误列表
Agent 过滤-a <name>只看指定 agent
限制数量-l N只显示最近 N 个 Run
指定日期YYYY-MM-DD默认今天

参数可组合:-s -a main-l 3 -a wairesearch

实时跟踪(-f)和高级模式(--advanced)需在 SSH 终端运行,
详见 references/advanced-mode.md

数据源

脚本有两种数据源,自动切换:

数据源路径需要配置精度
Debug 日志/tmp/openclaw/openclaw-YYYY-MM-DD.logdiagnostics.enabled: true精确 Run 边界
Session 文件~/.openclaw/agents/*/sessions/*.jsonl无需配置虚拟 Run(消息时间戳推算)

无 debug 日志时自动降级为 session 模式,核心指标(推理耗时、Token、工具统计)仍然准确。

输出内容

摘要统计

  • 模型调用次数、平均推理延迟、Token 吞吐量
  • 工具调用次数、成功率、总耗时
  • Thinking 统计(次数、平均深度)
  • Per-Agent 活动分布

Run 详情(非摘要模式)

  • 每个 Run 的时间线(推理段 + 工具调用段)
  • 推理耗时、输出 Token、吞吐速率
  • 工具调用参数摘要

错误列表

  • 最近 20 条错误,按时间倒序

使用指南

日常检查

# 快速了解今天的运行概况
bash scripts/byted-openclaw-diag.sh -s

性能排查

# 查看某天详细 Run 数据,找到慢查询
bash scripts/byted-openclaw-diag.sh 2026-03-19 -l 10

特定 Agent 分析

# 只看 waicode 的活动
bash scripts/byted-openclaw-diag.sh -a waicode -s

实时监控(SSH 终端)

# 需在 SSH 终端运行,不适合 Telegram/聊天
bash scripts/byted-openclaw-diag.sh -f
bash scripts/byted-openclaw-diag.sh -f --advanced

注意事项

  • 脚本依赖 python3(3.7+,使用 datetime.fromisoformat)和 bash
  • 高级模式(--advanced)会临时修改 openclaw.json 并重启 Gateway,退出时自动恢复
  • 无 Swap 的机器上并发多 Agent 时注意内存
  • 时间戳统一为 UTC 处理,不受本地时区影响

Related skills

FAQ

What data sources does byted-openclaw-diag use?

It reads OpenClaw debug logs for precise run boundaries, or falls back to session files, keeping core metrics accurate.

Does it support live tracking?

Yes, the -f live-tracking and --advanced modes run in an SSH terminal, not in a chat channel.

Debuggingmonitoring

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.