AgentSkillsCN

investigate-logs

在出现错误、排查调试问题或进行故障排除时,自动对日志进行深入分析。通过前端日志(LogViewer)和后端日志(每日日志文件)查找常见的错误模式。当用户反馈 Bug、系统报错,或功能未能按预期运行时,可随时调用此工具。

SKILL.md
--- frontmatter
name: investigate-logs
description: Automatically investigate logs when errors occur, debugging issues, or troubleshooting problems. Check frontend logs (LogViewer) and backend logs (daily log files) for error patterns. Use when user reports bugs, errors appear, or functionality isn't working as expected.

Investigate Logs

Automatically investigate PhotoClove logs to diagnose issues and errors.

When This Skill is Invoked

Claude automatically uses this skill when:

  • User reports an error or bug
  • Functionality isn't working as expected
  • Need to trace what happened during an operation
  • Debugging why something failed
  • User says "check the logs" or similar

Log Locations

Frontend Logs

Location: LogViewer component (accessible via Ctrl+Shift+L)

  • Stored in memory via LoggerService.js
  • Structured format: [LEVEL] Component:event_name - Message | data: {...}
  • Can be filtered by level, component, and searched

Daily Log Files (UTC dates):

bash
~/.local/share/photoclove/logs/photoclove-frontend-YYYY-MM-DD.log

Example: photoclove-frontend-2025-12-23.log (December 23, 2025 UTC)

Backend Logs

Location: Daily log files in the logs directory (UTC dates)

bash
~/.local/share/photoclove/logs/photoclove-YYYY-MM-DD.log

Example: photoclove-2025-12-23.log (December 23, 2025 UTC)

Format: Rust log:: macros with structured key=value pairs

code
[TIMESTAMP] [LEVEL] target: "component", "event; key1=value1; key2=value2"

Important: Log file names use UTC dates, not local timezone dates.

Investigation Process

  1. Identify the issue:

    • What operation failed?
    • What was the user doing?
    • When did it happen?
  2. Check most recent logs (log files use UTC dates):

    bash
    # Frontend logs (today UTC)
    tail -100 ~/.local/share/photoclove/logs/photoclove-frontend-$(date -u +%Y-%m-%d).log
    
    # Backend logs (today UTC)
    tail -100 ~/.local/share/photoclove/logs/photoclove-$(date -u +%Y-%m-%d).log
    
    # Or simply check the latest files
    tail -100 ~/.local/share/photoclove/logs/photoclove-frontend-*.log | tail -100
    tail -100 ~/.local/share/photoclove/logs/photoclove-*.log | grep -v frontend | tail -100
    
  3. Search for errors:

    bash
    # Find errors in frontend logs
    grep -i "error" ~/.local/share/photoclove/logs/photoclove-frontend-*.log | tail -20
    
    # Find errors in backend logs
    grep -i "error" ~/.local/share/photoclove/logs/photoclove-2*.log | tail -20
    
  4. Trace operation flow:

    • Look for correlation_id in logs to trace a request through frontend and backend
    • Follow the sequence of events leading to the error
    • Check what data was passed at each step
  5. Identify root cause:

    • What was the first error in the chain?
    • What state was the application in?
    • What data caused the issue?

Common Error Patterns

Frontend Errors

  • ReferenceError: Undefined variable or function
  • TypeError: Wrong type or null/undefined access
  • State errors: Incorrect state updates or missing state

Backend Errors

  • Database errors: SQL errors, table not found, constraint violations
  • File system errors: Permission denied, file not found
  • Tauri command errors: Command not found, invalid parameters

Integration Errors

  • 500 Internal Server Error: Backend error, check backend logs
  • No response: Command timeout or backend crash
  • Data mismatch: Frontend and backend data format mismatch

Log Analysis Tips

  1. Start from the error and work backwards
  2. Look for correlation_id to trace requests
  3. Check timestamps to understand timing
  4. Pay attention to data values in structured logs
  5. Compare with successful operations

Reporting Findings

After investigation, provide:

  • Root cause: What actually caused the error
  • Error location: File and line number if available
  • Data context: What data triggered the issue
  • Suggested fix: How to resolve the problem
  • Related logs: Key log entries showing the issue

Integration with PhotoClove

  • Frontend logs use LoggerService.js for structured logging
  • Backend logs use Rust log:: macros
  • All operations should have correlation IDs for tracing
  • LogViewer provides UI for browsing logs interactively