Detecting Harness Lockin

What this Claude Code skill does

Identifies and quantifies the compounding switching cost of an AI coding agent harness commitment before it becomes invisible. Use when evaluating whether to change AI coding agent platforms, when pricing a harness decision for a team or organization, or when auditing how much process infrastructure has accumulated around a current agent architecture. Common phrases: "how locked-in are we to this

Detecting Harness Lockin is a packaged skill for Claude Code: a SKILL.md plus references and a runnable test that drops into .claude/skills/, so your agent can run this workflow reliably instead of you re-explaining it each time. Install it via MCP, run it as a CLI, or fork it into your own.

Inside the SKILL.md

<h1>Detecting Harness Lock-in</h1>
<h2>The Lock-in Mechanism</h2>
<p>When a team adopts an AI coding agent, they do not just adopt a subscription.
They begin accumulating process infrastructure specifically shaped by that harness:</p>
<ul>
<li>Workflow automation built around the harness's abstractions (skill systems, context
  forking, sub-agent models)</li>
<li>Context files that encode project history in formats the harness expects</li>
<li>MCP connectors deployed and tuned for the harness's integration model</li>
<li>Team habits, verification steps, and review protocols designed around agent behavior</li>
<li>Documentation structured to be legible to that specific agent</li>
</ul>
<p>Each layer compounds. After six months, switching harnesses does not mean learning
new commands — it means rebuilding a layered chain of automation in an architecture
that may not support the same abstractions at all.</p>
<blockquote>
<p>"The organization's building workflows around these tools. They're not just adopting
a subscription. They're building institutional knowledge and process documentation
and verification protocols around a specific agent architecture."</p>
</blockquote>
<p>This is the lock-in nobody prices into their initial decision.</p>
<hr />
…