What is the Agent Issue Logger skill?

It is a specialized Claude Code skill that allows subagents to record errors, warnings, and events to a structured log file in the background.

What information is automatically added to each log entry?

The logger automatically adds a timestamp and the current CLAUDE_SESSION_ID to every entry for better traceability.

Where are the agent logs stored?

All entries are appended to a centralized file located at ~/.claude/logs/agent-issues.jsonl.

Does logging issues slow down my Claude session?

No, the skill is designed to run in the background with silent failures (2>/dev/null), ensuring your primary coding workflow remains fast and uninterrupted.

Agent Issue Logger

Name: Agent Issue Logger
Author: nathanvale

bynathanvale

•

分析と監視

Logs structured errors, warnings, and events from subagents to a centralized JSONL file without interrupting active workflows.

Agent Issue Logger is a high-performance, background utility designed for Claude Code subagents to record execution details, tool failures, and degraded results. By leveraging a fire-and-forget mechanism, it ensures that developers can track agent health and debug complex automated workflows through a centralized log file (~/.claude/logs/agent-issues.jsonl) without stalling the main execution thread. Part of the SideQuest suite, it is optimized for maintaining developer hyperfocus by handling telemetry silently and efficiently.

主な機能

01Support for multiple severity levels including error, warning, and info

02Batching support to log multiple related issues in a single call

033 GitHub stars

04Automatic timestamping and session ID enrichment

05Structured JSONL output for easy parsing and monitoring

06Fire-and-forget background logging via run_in_background

ユースケース

01Debugging tool failures or timeouts in background agent processes

02Collecting execution telemetry for long-running developer workflows

03Monitoring subagent reliability during complex multi-step autonomous tasks

主な機能

01Support for multiple severity levels including error, warning, and info

02Batching support to log multiple related issues in a single call

033 GitHub stars

04Automatic timestamping and session ID enrichment

05Structured JSONL output for easy parsing and monitoring

06Fire-and-forget background logging via run_in_background

ユースケース

01Debugging tool failures or timeouts in background agent processes

02Collecting execution telemetry for long-running developer workflows

03Monitoring subagent reliability during complex multi-step autonomous tasks