development-process-review

Analyzes Claude Code session transcripts to produce an agile development-process retrospective report.

Updated May 3, 2017
One-click install
npx skills add https://github.com/arnonmoscona/dot_files --skill development-process-review-arnonmoscona
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: development-process-review
Source: https://github.com/arnonmoscona/dot_files/tree/main/wsl/.claude/skills/development-process-review
Command: npx skills add https://github.com/arnonmoscona/dot_files --skill development-process-review-arnonmoscona

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve? Reviewing weeks of Claude Code session transcripts by hand is impractical, so teams lack evidence-based feedback on how their human-plus-AI development workflow actually performed and what guidance in CLAUDE.md should change. ## Core Features & Use Cases - Transcript Digestion: Condenses raw JSONL session transcripts into per-day Markdown digests via scripts/digest.py, stripping low-signal tool noise while keeping user and assistant text. - Parallel Batch Analysis: Fans out digested batches to parallel background agents, then merges and deduplicates findings into one report. - Two-Sided Retrospective: Evaluates both Claude's behavior and the human supervisor's behavior across continue/stop/improve categories, ending with prioritized CLAUDE.md edit candidates. - Use Case: After a three-week development stretch, ask for a process review since a given date and receive a structured retrospective covering both team members plus concrete guidance changes. ## Quick Start Ask the assistant to run a development process review since a specific date, for example: review our development process since 6/23/2026.

Frequently Asked Questions about development-process-review

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

FAQPage Schema
How do I review my Claude Code development process over past sessions?

Invoke the review with a since-date argument, such as "since 6/23/2026". The skill locates session JSONL transcripts, digests them by day, analyzes batches with parallel agents, and merges findings into a structured retrospective report.

How does the review handle very large transcript volumes?

A digest script strips low-signal tool calls like Read, Glob, and Grep while keeping user and assistant text, cutting volume by over 90 percent. Daily digests are then grouped into 300-500KB batches analyzed by parallel background agents.

Does the retrospective only evaluate the AI assistant's behavior?

No, it explicitly evaluates both team members. Each continue/stop/improve category has subsections for Claude's behavior, the human supervisor's direction and review habits, and team interaction patterns.

Where does the skill find Claude Code session transcripts?

Transcripts are read from ~/.claude/projects/<cwd-with-slashes-replaced-by-dashes>/ as one JSONL file per session. Files are filtered by checking whether their first and last timestamps overlap the requested date range.

What is the final output of the development process review?

The output is a Markdown artifact report organized by continue/stop/improve categories across Claude, supervisor, and team dimensions. It closes with a prioritized list of concrete CLAUDE.md edit candidates for both parties.