AgentStack
SKILL verified MIT Self-run

Openclaw Logs

skill-jincai-openclaw-skills-openclaw-logs · by jincai

查询和分析 OpenClaw gateway 日志。当用户提到「查日志」「openclaw 发生了什么」「openclaw 日志」「gateway 日志」「查看最近的错误」「openclaw 最近怎么了」「openclaw 报错」或询问 OpenClaw 在某个时间段内的运行情况时,使用此技能。支持时间范围查询、按类别过滤(error/restart/telegram/plugin/acp/config 等)、自动去噪、时区转换,输出结构化的事件时间线。

No reviews yet
0 installs
19 views
0.0% view→install

Install

$ agentstack add skill-jincai-openclaw-skills-openclaw-logs

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access Used
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

Are you the author of Openclaw Logs? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

OpenClaw Log Query

查询 OpenClaw gateway 日志,生成结构化的事件时间线。

Quick Start

运行脚本获取日志数据,然后基于输出为用户分析:

# 查询指定时间段(北京时间)
python3 ~/.claude/skills/openclaw-logs/scripts/query_logs.py --from "19:20" --to "20:00"

# 查询最近 N 分钟/小时
python3 ~/.claude/skills/openclaw-logs/scripts/query_logs.py --last 30m
python3 ~/.claude/skills/openclaw-logs/scripts/query_logs.py --last 2h

# 只看错误
python3 ~/.claude/skills/openclaw-logs/scripts/query_logs.py --from "19:00" --to "20:00" --category error

# 只看重启事件
python3 ~/.claude/skills/openclaw-logs/scripts/query_logs.py --last 1h --category restart

# 指定日期
python3 ~/.claude/skills/openclaw-logs/scripts/query_logs.py --from "2026-03-03 22:00" --to "2026-03-04 02:00"

# 详细模式(含 cron 心跳等噪声)
python3 ~/.claude/skills/openclaw-logs/scripts/query_logs.py --last 1h --verbose

Parameters

| Parameter | Description | Example | |-----------|-------------|---------| | --from | Start time (HH:MM or YYYY-MM-DD HH:MM), Beijing time | "19:20", "2026-03-04 19:20" | | --to | End time, Beijing time | "20:00" | | --last | Relative time window from now | 30m, 2h | | --date | Override date for HH:MM times (default: today) | 2026-03-03 | | --category / -c | Filter by event category | error, restart, plugin, telegram, acp, config, agent, gateway | | --verbose / -v | Include normally-filtered noise (cron heartbeats etc) | |

Categories

The script auto-categorizes each log entry:

  • ERROR - Failures, rate limits, authentication errors
  • RESTART - SIGTERM/SIGUSR1, draining, shutdown/startup
  • PLUGIN - Plugin install, load, acpx backend
  • TELEGRAM - Message sends, bot command issues
  • ACP - ACP identity reconcile, agent dispatch
  • CONFIG - Config changes, reload, provider updates
  • WS - WebSocket connections (webchat UI)
  • HEALTH - Health monitor checks and restarts
  • AGENT - Agent model info, embedded run results
  • GATEWAY - Listener, canvas, heartbeat, browser control
  • CRON - Cron job triggers

Log Sources

The script reads from multiple sources and deduplicates:

  1. ~/.openclaw/logs/gateway.log - Plain text, mixed UTC/+08:00 timestamps
  2. ~/.openclaw/logs/gateway.err.log - Plain text, errors and warnings
  3. /tmp/openclaw/openclaw-YYYY-MM-DD.log - Structured JSON (one file per day)

All timestamps are normalized to Beijing time (UTC+8) in the output.

Workflow

  1. Run the script with the user's time range
  2. Read the output - it includes a Summary (category counts), Errors (highlighted), and Timeline (chronological)
  3. Provide analysis to the user: what happened, what's concerning, and any suggested actions
  4. If the user asks about a specific issue, use --category to drill down

When the user's question is vague (like "openclaw 怎么了"), start with --last 30m to see recent activity, then narrow down based on what you find.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.