agent-harness-optimizer

name	agent-harness-optimizer
description	Agent harness optimization patterns for token efficiency, memory persistence, session management, and cross-harness parity. Use when optimizing agent performance, reducing token costs,.
domain	core
author	oyi77
license	Apache-2.0
subdomain	core-platform
tags	["token-optimization","memory-persistence","session-management","hooks","cross-harness","performance"]
persona	{"name":"NanoClaw","title":"Harness Performance Architect","expertise":["Token Optimization","Hook Systems","Memory Persistence","Cross-Harness Engineering","Session Lifecycle Management"],"philosophy":"Every token spent must earn its place. Optimize the harness, not just the model."}
version	1.0.0

Overview

Agent harness optimization is the practice of tuning the runtime environment that surrounds an AI agent -- model selection, prompt structure, hook configuration, memory persistence, and session management -- to maximize output quality while minimizing token cost and latency. Derived from real-world patterns across 10+ months of daily agentic work, these techniques apply to any harness: Claude Code, Cursor, OpenCode, Codex, Gemini, and beyond.

Anti-Rationalization Table

Rationalization	Reality
"I'll figure it out as I go"	A structured approach saves time and reduces errors. Follow the workflow in this skill rather than improvising.
"I already know this topic"	Familiarity breeds shortcuts. Use the checklist to verify you haven't missed critical steps.
"This doesn't apply to my situation"	The patterns here generalize across contexts. Adapt, don't skip — the underlying principles hold.
"One more tool will fix it"	Adding complexity rarely solves process gaps. Master the core workflow first.

When to Use

Trigger phrases:

• "agent harness optimizer"

• "Token costs are rising faster than output quality"

• "Agents lose context between sessions or after compaction"

• "Hook scripts are slow, brittle, or produce noisy output"

• Token costs are rising faster than output quality

• Agents lose context between sessions or after compaction

• Hook scripts are slow, brittle, or produce noisy output

• You need the same agent behavior across multiple AI coding harnesses

• Session history grows unwieldy and needs structured management

• Background processes are eating into the main context window

• You want to set up continuous learning from session patterns

When NOT to Use

• Task is outside your authorization scope
• You need to implement controls (use implementing-* skills)
• Task is about analysis, not action (use analyzing-* skills)
• You don't have access to target systems
• Task requires compliance expertise (consult professionals)
• Task is about defense, not offense (use defensive skills)

Process

1. Gather requirements and constraints from the user