Stories by shad42

Automating the engineering work that isn't your product

How we know if our agent is right

Not update your dependencies in 2026

Supply chain attacks don't wait for CVEs

We built an agent for CI despite using Claude Code for everything else

We built an agent specialized for CI (using the 3 Claude models)

How we know if our agent is right

How we know if our agent is right

The agent harness belongs outside the sandbox

We decreased our LLM costs with Opus

Multi-player agents don't fit in the sandbox

We built our AI agent, for analyzing CI logs

Same LLM, different agent: a CI debugger built on Claude

Agent Harness: Inside vs. Outside the Sandbox

Same LLM but different output: we built a CI specialist

We upgraded our agent to Opus and our costs went down

Same LLM, Different Agent: What Changes When You Specialize for CI

We decreased our LLM costs by switching to Opus

What CI looks like at a 100-person team (PostHog)

We upgraded to a frontier model and our costs went down

1