digest

Converts CDash test reports to a unified format for easy analysis.

Updated Jan 8, 2026
One-click install
npx skills add https://github.com/thekevinscott/digest --skill digest-thekevinscott
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: digest
Source: https://github.com/thekevinscott/digest/tree/main
Command: npx skills add https://github.com/thekevinscott/digest --skill digest-thekevinscott

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps teams quickly convert Claude Code transcripts into actionable artifact notes, reducing manual summarization and enabling faster decision making.

Core Features & Use Cases

  • Transcript extraction: parse user and assistant messages, including tool uses and results, into clean, labeled lines.
  • Insight generation: categorize findings into opportunities, loose ends, backlog, and learnings for easy review.
  • Use Case: Example: process a conversation log to produce digest notes that can be stored in notes/digest/.

Quick Start

Use the digest skill to view unprocessed transcripts or extract insights from a session using the Digest API or CLI. For example, run digest list to see transcripts with unprocessed lines or digest extract <session-id> to fetch the next batch of content.

Frequently Asked Questions about digest

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

FAQPage Schema
How do I extract insights from Claude Code transcripts automatically?

You can extract insights from Claude Code transcripts using the digest CLI by running digest list to find unprocessed sessions and digest extract <session-id> to fetch the next batch of content. This parses user and assistant messages, including tool uses, into categorized artifact notes.

What is the best way to turn conversation logs into actionable artifact notes?

Turning conversation logs into actionable artifact notes involves parsing raw transcript sessions to extract structured insights. Digest categorizes findings into opportunities, loose ends, backlog, and learnings, reducing manual summarization and enabling faster review of multi-session files.

Can I process multi-session transcript files and paginate through the results?

Processing multi-session transcript files is supported by digest, which handles single or multi-session logs and returns a next_cursor for pagination. It uses per-extraction budgets and line overlap to manage large files, exposing list, extract, mark, and note operations via its Python API and CLI.

Does transcript extraction work with tool interactions and raw session messages?

Yes, transcript extraction works with raw session messages and tool interactions. Digest parses Claude Code transcripts containing user and assistant messages alongside tool uses and tool results, converting them into clean, labeled lines for categorized artifact note generation.

Do I need a Python environment to use the transcript analysis API?

A Python environment is needed to use the transcript analysis API, as digest exposes list, extract, mark, and note operations via a Python API. Alternatively, the CLI provides direct access to run digest list or digest extract <session-id> without writing Python code.