Getting started

This is a library of expertise, packaged as sixteen independently installable departments. You install the ones you need, and the relevant specialist loads itself when you ask something in its territory.

It runs in Claude Code and in ChatGPT and Codex. The skills are the same files in both — only the manifests differ, and both sets are generated from the same tree, so a fix reaches both at once.

Nothing here changes your project. Installing a department adds skills to the assistant — it does not write to your repository, add dependencies, or run anything on its own.

Install

Claude Code

Add the marketplace once:

/plugin marketplace add cbrock84/headcount

Then install departments one at a time:

/plugin install it-operations@headcount
/plugin install security@headcount

/plugin on its own opens the plugin menu, where installed plugins can be reviewed and managed.

ChatGPT and Codex

The same repository. Add it as a plugin marketplace — the manifest ChatGPT reads is at .agents/plugins/marketplace.json — or copy the department you want into .agents/skills/ in your own project.

One thing worth knowing on that side: skill descriptions share a context budget of roughly eight thousand characters, and all sixteen departments together come to about fifty thousand. Installing a few departments is not just tidier there, it is the only thing that works.

Install what you will use, not everything. Sixteen departments is a lot of surface, and a smaller set produces sharper triggering. You can add more at any point.

Which departments to install first

Start from what you actually spend your week on rather than from the org chart.

If you…	Start with
Run IT or infrastructure for a company	it-operations, security, operations
Run a small company or a whole function	executive, finance, operations
Own revenue or go-to-market	revenue, marketing, demand-generation
Build product	product, technology, data-analytics
Run delivery or a PMO	pmo, operations
Own people or hiring	people, legal-risk
Have customers to keep	customer-experience, revenue

Add security and legal-risk early regardless of what else you install. They are the two that tell you something you did not want to hear, which is the whole reason to have them.

Three ways to use it

1. Just ask. Skills load themselves when a request matches. This is the normal path and the one to reach for first.

"Our margins slipped this quarter and nobody can tell me why."

2. Name the skill when you want a specific lens rather than the one that would trigger. In Claude Code:

/finance:cost-accounting

Elsewhere, name it in the sentence — "use the cost accounting skill" — since the slash form is Claude Code's. Skills are addressed as department:skill throughout, so names never collide.

3. Delegate a whole department. Each department ships an agent charter in .claude/agents/, so it can be handed a body of work as a subagent with its own exclusive write surface. See executive:agent-hierarchy for the method, and why the split is by write surface rather than by topic.

What to expect back

A skill is not a template that fills itself in. Most will start by asking you the two or three things that decide the answer — what the unit is, what the comparison is, who owns the consequence — because those are the questions that get skipped and are the reason the usual answer is wrong.

Expect to be told what the library cannot determine. A skill that names its limits is doing the job; one that answers everything confidently is the failure mode this catalog is built against.

Reviewer-class departments behave differently

security and legal-risk review what other departments commit to. They hold no write surface, they report to the chief executive rather than into the function they review, and their blocking findings are not overrulable by the department under review.

In practice: when you ask a revenue question and the security review says the access model cannot support the control the deal requires, that is not a trade-off to price against the deal. The date moves or the control gets built. This is deliberate, and it is why those two are worth installing even when they are not your job.

Some skills cite outside authorities

Where a question is settled by a regulator, a standards body or primary law, the skill carries a list of those sources at references/sources.md inside it, and reads them while answering rather than recalling them. Tax, employment, security controls, privacy and education standards are covered so far. You will see them cited in the answer.

Two things worth knowing about how that behaves.

The agent may fetch a source, and a fetched page is input rather than instruction. If a page appears to be telling the assistant to do something, that is content on someone else's website, not a request from you. Treat an answer that changed direction after a fetch with the same suspicion you would treat any other surprise.

A cited source is not a quoted one, on purpose. Most of what a professional must cite is not open — ISO standards are sold, SANS papers are copyrighted, the FASB Codification needs an account. Each entry records what may be done with it, and where the answer is "read and cite only" the skill will name the standard and its clause rather than reproduce the text. That is the correct behavior, not an omission. The full index is here.

What this is not

Several skills carry an explicit disclaimer, and it is not boilerplate. Employment classification, tax, financing terms, contract positions, privacy obligations and regulatory filings are matters where the facts and the jurisdiction decide the answer. These skills structure the question and tell you what to ask — they are not a substitute for qualified counsel, and they say so.

Seeing what is in it

The live org chart is searchable across every skill in every department, generated from the repository so it cannot drift from what actually ships.

The source catalog lists every outside authority the skills check against, which skills use each, and what you may do with it.

Worked situations show what happens when a problem crosses departments — what engages, in what order, what comes back, and where the library has nothing useful to say.

Getting help

Questions, gaps and disagreements belong in Discussions. A situation the catalog handles badly is the most useful thing you can report, more than a missing skill — the gap is usually that a skill exists and fails to trigger, or triggers and answers the wrong question.