Skip to content

Open source

ddogctl

A CLI for the Datadog API, built end to end with Claude Code as the method.

Role
MIT
Stack
Python 3.10+ · Datadog API · Claude Code · PyPI

Source on GitHub · Package on PyPI

The gap

Investigating a production incident through the Datadog web UI means a lot of tab-switching: a trace here, the correlated logs there, the slow query in a third place. I wanted that flow in a terminal, close to where I was already working, so I built it as an internal tool at Kojo to speed up production investigation, then genericised it and open-sourced it under MIT.

Dogshell already exists and covers a different surface; the ddogctl README compares the two directly.

What it does

  • APM trace search
  • Log querying with trace correlation
  • Database monitoring — slow queries and execution plans
  • Monitors
  • Cross-signal investigation workflows that stitch the above together

Eight releases, v1.0.0 through v2.1.0, between February and May 2026. 74 commits.

Why it is here

Not because of its reach — it is a small tool with a small audience. It is here because it is the one piece of my AI-assisted development argument that is public and inspectable rather than described.

I built it end to end with Claude Code as the development workflow: initial design, implementation, packaging, releases. Then I maintained it for three months and eight releases. Everything about that is verifiable from the commit history and the PyPI release timeline, which is more than I can offer for work that happened inside private repositories.

← All projects