Intellectual Frontiers Press Fieldbooks

Building the Workforce

Setting Up the Repository and Building the Inputs for an AI Workforce That Finds Who Is Ready to Act

Prompts teach each step, and skills and MCPs run it.

Quarry Lane Health is an 11-person company that sells a site readiness setup and a platform subscription to independent clinical research sites. Its engineer starts a customer workforce in one Git repository from thirteen empty specification stubs in January, and turns them into rules the workforce reads before it contacts anyone, each with labeled cases a test runs on every change. The next volume runs that workforce until it reaches Run in September.

Forthcoming. The book is in review.

Build the files a readiness workforce reads before it contacts anyone, and test every one.

Every step appears first as a prompt, because a prompt shows what the step asks of the model. Then the book turns each repeated step into a skill, an MCP server, or a script with labeled test cases, because nobody should run a workforce by typing prompts in. Tools, versions, and law are stated as of the edition, and the online appendix keeps the dated tables current.

What you will be able to do

• start the repository, its specification stubs, and the review rules for each role;
• write the outcome and the counterparty profile as files a test can check;
• build the actor map, signals, readiness model, and access policy with labeled cases;
• specify each motion and its response surface, and record only what email and web tracking can see; and
• turn each repeated step into a tested skill, MCP server, or script, or find one that already exists.