Qmd Sessions
Turn raw Claude Code session logs into markdown your qmd index can search so you can find past prompts, fixes, and decisions without scrolling JSONL.
Overview
qmd-sessions is an agent skill most often used in Build / agent-tooling (also Operate / iterate) that converts Claude Code session transcripts into markdown files ready for qmd indexing.
Install
npx skills add https://github.com/wbelk/claude-qmd-sessions --skill SKILL.mdWhat is this skill?
- Converts Claude Code session transcripts into clean markdown suited for qmd indexing
- Makes long agent threads searchable alongside docs and notes in a local knowledge base
- Pairs with qmd-style “query my docs” workflows instead of manual copy-paste from session files
- Lightweight pipeline: transcript in → markdown files out for offline search
- No external API—works on exported or local session storage you already have
Adoption & trust: 46 GitHub stars.
What problem does it solve?
You have weeks of Claude Code sessions in transcript form but no fast way to search decisions, commands, or fixes across them.
Who is it for?
Mac or Linux solo builders who use Claude Code heavily, run qmd on a markdown corpus, and want session history in that corpus.
Skip if: Teams that need centralized cloud session analytics, real-time multi-user chat search, or workflows with no local transcript files to convert.
When should I use this skill?
You have Claude Code session transcripts and want them converted to markdown for qmd or local full-text search.
What do I get? / Deliverables
You get markdown session files your qmd index can query so past agent work becomes discoverable like the rest of your repo docs.
- Markdown files derived from session transcripts
- A corpus layout suitable for qmd indexing
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Session export is part of the agent toolchain—canonical shelf is Build because it operates on Claude Code artifacts, not on production runtime. Agent-tooling is where skills that extend Claude Code workflows (transcripts, indexing, retrieval) belong in Prism’s browse graph.
Where it fits
After a long Claude Code refactor, export sessions to markdown so qmd can answer “what did we change in auth?”
When a regression appears, search indexed sessions for the original fix discussion instead of re-debugging from scratch.
Before merge, pull cited session snippets from indexed markdown to document rationale in the PR.
Build an internal changelog from session exports when you skipped formal docs during a sprint.
How it compares
Local transcript-to-markdown prep for search indexing—not a hosted agent analytics or MCP memory server.
Common Questions / FAQ
Who is qmd-sessions for?
Solo and indie builders who use Claude Code and qmd (or similar markdown search) and want past coding sessions findable without manual export gymnastics.
When should I use qmd-sessions?
During Build when you archive agent work for docs; during Operate when you iterate and need to recall how a bug was fixed; anytime before writing specs or reviews that should cite real session context.
Is qmd-sessions safe to install?
It processes local session transcripts on your machine—review what paths it reads and writes. Check the Security Audits panel on this Prism page before running it on sensitive repos.
SKILL.md
READMESKILL.md - Qmd Sessions
Convert Claude Code session transcripts to searchable markdown files for qmd indexing # qmd-sessions { "name": "qmd-sessions", "description": "Convert Claude Code session transcripts to searchable markdown files for qmd indexing" }