history-insight

Extract structured insights from Claude Code session JSONL logs.

Updated Apr 9, 2026
One-click install
npx skills add https://github.com/treejh/ai-agent-skills --skill history-insight-treejh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: history-insight
Source: https://github.com/treejh/ai-agent-skills/tree/main/plugins/session-wrap/skills/history-insight
Command: npx skills add https://github.com/treejh/ai-agent-skills --skill history-insight-treejh

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires jq, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Claude Code session files are large and noisy, making it hard to find, extract, and summarize the actual conversation content; this Skill locates session JSONL files, filters out metadata and tool traces, and surfaces concise insights and counts for decisions, action items, and patterns.

Core Features & Use Cases

  • Session discovery: Find session JSONL files for the current project or across all projects using configurable scope and date filters.
  • Message extraction: Keep only user messages and assistant text responses while discarding file snapshots, queue logs, thinking traces, and tool_use entries.
  • Batch processing & scaling: Use a cache and split-and-parallelize pipeline for many or very large session files, then merge batch analyses.
  • Use case: Capture today's conversation to generate an engineering devlog, extract decisions and action items from past discussions, or audit session content for knowledge transfer.

Quick Start

Capture and summarize all session messages for the current project and extract key decisions, action items, and message counts.

Frequently Asked Questions about history-insight

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract conversation content from Claude Code JSONL session files?

This Skill extracts conversation content from Claude Code JSONL session files by filtering out metadata, tool traces, and file snapshots, retaining only user and assistant text messages for structured summarization and pattern detection.

What is the best way to summarize Claude Code session history for a devlog?

Summarizing Claude Code session history for a devlog involves processing JSONL files to locate user and assistant messages, then aggregating the filtered text to extract key decisions, action items, and message counts.

Does this session analysis tool require jq to process JSONL logs?

Yes, this session analysis tool requires jq as a dependency to parse and filter local Claude Code JSONL logs, invoking the extraction script to aggregate user and assistant text messages for batch processing and summarization.

Can I batch process large JSONL session files across multiple projects?

Yes, you can batch process large JSONL session files across multiple projects using a cache and split-and-parallelize pipeline that preprocesses large files and merges batch analyses for comprehensive conversation insights.

How do I filter out tool_use entries and thinking traces from Claude Code conversations?

Filtering out tool_use entries and thinking traces from Claude Code conversations requires applying message filtering during JSONL extraction to keep only user messages and assistant text responses while discarding queue logs and file snapshots.