How to Size Tasks for AI Coding Agents

Mon, 27 Apr 2026 00:00:00 +0000

Getting task scope right is the difference between an agent that ships clean code on the first try and one that spirals into corrections, context exhaustion, and wasted tokens. Most people size tasks by gut feel — “that seems about right” — but the actual constraint is measurable, and the research on what makes a reviewable unit of work is well-established.

The Real Constraint: Context, Not Lines of Code

People instinctively think about task size in terms of lines of code or number of files changed. Those are secondary proxies. The actual limiter is how much context the agent must consume — reading files, exploring the codebase, running commands, processing outputs — before it can do the work.

The Anatomy of a Perfect AI Agent Task

Mon, 20 Apr 2026 00:00:00 +0000

A well-crafted task for an AI coding agent is essentially context engineering — you’re deliberately curating the minimum set of information the agent needs to produce the right output on the first try. Rather than pre-loading everything up front, the best approach combines focused instructions with enough pointers that the agent can pull in additional context just-in-time as it works (Anthropic — Effective Context Engineering ). Below is a breakdown of every element that matters, why it matters, and a full example at the end that ties it all together.

Blog on John Young

How to Size Tasks for AI Coding Agents

The Real Constraint: Context, Not Lines of Code

The Anatomy of a Perfect AI Agent Task