Query Claude Code session transcripts in JSONL files for project history.
npx skills add https://github.com/dundas/uhr --skill transcript-query
Please help me install this Agent Skill. Skill: Transcript Query Source: https://github.com/dundas/uhr/tree/main/.claude/skills/transcript-query Command: npx skills add https://github.com/dundas/uhr --skill transcript-query
This Skill helps you quickly recall and understand past project decisions, work, and context by searching through your Claude Code session transcripts.
Use the transcript query skill to find out what work was done before the 'daemon' topic.
High-intent search queries and answers about installing and using this skill.
Yes, you can find decision-making rationale within session transcripts by performing keyword-based queries against stored JSONL files to extract the historical context and reasoning behind specific project approaches.
Filtering JSONL transcript files by date allows you to reconstruct the timeline of work on a particular feature, helping you see when specific files were modified and what issues were encountered.
When you need to understand why specific approaches were chosen or what errors occurred with a feature, querying session transcripts helps retrieve the exact decision-making rationale and error logs from your past work.
You can track file modification history by searching session transcripts to see when specific files were last modified, retrieving chronological context directly from stored JSONL session logs.
Limitations may arise when querying large JSONL session logs, as reconstructing error history relies on the available session data and the specificity of your keyword-based search queries to accurately retrieve past logs.
maxvaegaParse local JSONL transcript logs to estimate pay-per-token API costs in USD and EUR.
broomvaDocument project decisions with context, alternatives, and rationale in Markdown.
xlelibertyGenerate structured release notes and lint commits using Conventional Commits standards.
chavangorakh1999Analyzes Node.js app logs for errors and traces requests via Swim-lane IDs.
