Inside Coding Agents — Coding Agent Harness Architecture
vpromise.github.io · 502 words · saved by 1 readers
Model ≠ Agent. Learn the agent harness around the model.
OPEN TEXTBOOK · ATLAS · LABORATORY Open the agent shell.See the real system. The model is only the engine. Study the harness that lets a coding agent read code, call tools, enforce permissions, manage context, and finish reliably. 13 runnable lessons0 API keys required11 replayable events 01Intentuser.message 02Modelrequest / response 03Policyvalidate / authorize 04Tooleffect / result 05Contextappend / prune HARNESSowns the loop RUNNABLEVERSIONEDEVIDENCE-BACKEDREPRODUCIBLEBEGINNER → RESEARCHER ONE CONTENT GRAPH · FOUR VIEWS Not a pile of links. A knowledge infrastructure.…
saved by
related reading
- Building a harness for open modelsparthsareen.com
- Agent Leaderboards · Which tools coding agents choose · Armaturearmature.tech
- Building reliable AI agents · parth sareenparthsareen.com
- A Guide to Claude Code 2.0 and getting better at using coding agents – sankalp's blogsankalp.bearblog.dev
- Arjun Virkarjunvirk.com
- Effective harnesses for long-running agents \ Anthropicanthropic.com
- What I learned building an opinionated and minimal coding agentmariozechner.at
- GitHub - shareAI-lab/learn-claude-code: Bash is all you need - A nano claude code–like 「agent harness」, built from 0 to 1github.com
- Welcome to Learn Harness Engineeringwalkinglabs.github.io
- Why We Built Our Own Background Agentbuilders.ramp.com
- Designing agentic loopssimonwillison.net
- Building Effective AI Agents \ Anthropicanthropic.com