# Design exercise: Coding Agent Systems

> Repository context, planning, file edits, tools, tests, sandboxes, review loops, and patch delivery. Implement designOutline() returning non-empty values for: repoContext, editLoop, verification. Each value must name a concrete mechanism or decision.

Canonical: https://learn.significanthobbies.com/curriculum/drills/practice-coding-agent-systems

Repository context, planning, file edits, tools, tests, sandboxes, review loops, and patch delivery. Implement designOutline() returning non-empty values for: repoContext, editLoop, verification. Each value must name a concrete mechanism or decision.

## The mechanism

A coding agent is a repository-aware control loop. It needs scoped context, reversible edits, executable verification, and a clear handoff boundary.

## Expected evidence

A design outline with repoContext, editLoop, verification plus an explicit failure mode or trade-off.

## Hints

- Name the failure mode before selecting the mechanism.

## Sources and practice

- [SWE-agent](https://swe-agent.com/)
- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://arxiv.org/abs/2310.06770)
- [OpenHands: An Open Platform for AI Software Developers](https://arxiv.org/abs/2407.16741)
- [Building Effective AI Agents (Anthropic Engineering)](https://www.anthropic.com/engineering/building-effective-agents)
- [Study Coding Agent Systems](https://learn.significanthobbies.com/curriculum/concepts/coding-agent-systems)
- [Try the interactive drill](https://learn.significanthobbies.com/drills/practice-coding-agent-systems)
- [All practice briefs](https://learn.significanthobbies.com/curriculum/drills/)