Agent Nirzabari 0 1 0

Coding Agent Harness Architecture

Overview

An LLM is neither a coding agent nor a product. A coding agent product is a text generation model coupled with a harness — a runtime that repeatedly builds context, calls the model, executes tool calls, persists state, renders to the user, recovers from failures, and loops back. This guide covers the 7-layer architecture of production-grade coding agents, drawn from analysis of real codebases including Codex (OpenAI), OpenCode, Claude Code, and Cursor.

The gap between a simple loop like while :; do cat PROMPT.md | claude-code; done and what Codex or Claude Code actually ship is product engineering. That gap — context management, safety boundaries, persistence, recovery, and multi-client support — is the subject of this skill.

When to Use

• Designing agentic systems that go beyond simple prompt loops
• Analyzing agent architectures across different products (Codex, OpenCode, Claude Code, Cursor)
• Implementing harness patterns for context building, tool orchestration, or safety
• Studying production-grade coding agent implementations and their trade-offs
• Understanding the LLM provider contract (streaming events, tool calling, multi-turn state)
• Building scalable autonomous coding systems with parallelization and specialization

Core Concepts: The 7 Layers

A coding agent harness consists of 7 layers:

1. Agent Loop — Conversation turns: system prompt → user → model → tools → model → ... Each turn can include hundreds of tool calls. Context management is the agent's core responsibility.
2. Context Building — Gathering relevant data: which files are relevant, what's been done, style conventions. Context engineering is UX engineering — the product decides what the model sees and when.
3. Tooling Systems — Tool registry with argument schemas, multi-modal support, and parameters. Examples: shell, file edit, code search, browser automation, screenshot analysis.
4. Safety — Allowlists/denylists, sandboxing, snapshotting, recovery. Safety is architecture — approvals, policies, sandboxes, and undo.
5. Replay / Persistence — Forking chats and environments for debugging. Systems are made of turns, tool calls, diffs, approvals, and events that must be restorable.
6. Client Surface (TUI / Web / IDE) — How the user interacts: TUI (OpenCode, Claude Code, Codex), IDE (Cursor, Antigravity, Replit Agent), or Web (Bolt.new, v0, Lovable).
7. Extensibility — MCP for tool connectivity, AGENTS.md for repo-specific instructions, Skills for Anthropic's convention, and Open Responses for provider-agnostic API shape.

Capability Jumps

Four major jumps in AI capability from a user's perspective:

1. GPT-3.5 (ChatGPT, November 2022) — the leap was the product itself, not just the model
2. GPT-4 (Spring 2023)
3. Reasoning models (o1-preview, then o3 in Spring 2025)
4. Actually useful agentic systems (late 2025, strong reasoning models paired with solid harnesses)

Advanced Topics

Harness Fundamentals: The simplest agent loop, ralph wiggum pattern, and why it breaks at scale → [Harness Fundamentals](reference/01-harness-fundamentals.md)

Context Engineering: Prompt architecture, AGENTS.md patterns, OpenAI's lessons on context management → [Context Engineering](reference/02-context-engineering.md)

Tooling Systems: Tool registries, compiled vs dynamic tool handlers, Codex orchestrator vs OpenCode plugins → [Tooling Systems](reference/03-tooling-systems.md)

Safety and Persistence: Allowlists, sandboxing, AST-based command analysis, replay architecture → [Safety and Persistence](reference/04-safety-and-persistence.md)

Client Surfaces and Extensibility: TUI vs Web vs IDE patterns, MCP integration, Skills, plugin systems → [Client Surfaces and Extensibility](reference/05-client-surfaces-and-extensibility.md)

LLM Provider Contract: Streaming events (SSE), three-phase tool call lifecycle, multi-turn thread state, portability → [LLM Provider Contract](reference/06-llm-provider-contract.md)

Harness Deep Dive: Codex (Rust monolith) vs OpenCode (TS/Bun control-plane) architecture comparison across all 7 layers → [Harness Deep Dive](reference/07-harness-deep-dive.md)

Scaling Case Study: Anthropic's C compiler case study — parallel Claude agents, git-based coordination, lessons learned → [Scaling Case Study](reference/08-scaling-case-study.md)

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: tangledgroup
• Source: tangledgroup/tangled-skills
• License: MIT

Install and usage instructions live in the source repository linked above.