Chase AI Skills — AI skill for Claude Code

Chase AI+ Skills

Claude Code skills for the [Chase AI+](https://www.skool.com/) community. This repo is a **Claude Code plugin marketplace** — install once, get every skill, update with one command.

Access to this repo comes through the Chase AI+ classroom. Keep the link inside the community.

---

Install

In Claude Code:

/plugin marketplace add cth9191/chase-ai-skills
/plugin install chase-ai@chase-ai-plus

To pull updates later:

/plugin marketplace update chase-ai-plus

**Prefer a URL?** You can point Claude Code straight at the repo instead of the `owner/repo` shorthand — same result:

/plugin marketplace add https://github.com/cth9191/chase-ai-skills

…or just drop the repo URL into Claude Code and ask it to add the marketplace — it'll wire it up for you.

---

What's inside

`loop-engineer`

An interview-driven **loop-engineering architect**. You bring a task you want to automate; it does the hard part — the *decisions*, not just the code — then builds and wires the whole thing.

What happens when you run it

flowchart TD
  A["You: describe a recurring task
('every week I want Claude to…')"] --> B["1 · Diagnose
where it sits on the ladder"]
  B --> C["2 · Interview
fast, one question at a time,
each with a recommended answer"]
  C --> D["3 · Build the full loop
skill · trigger · state · verify · stop · README"]
  D --> E["4 · Test run
see one real output first"]
  E --> F{Happy with it?}
  F -->|yes| G["Wire it up
cron · Claude routine · OS task · manual"]
  G --> H["✅ Live & hands-off
(you only show up at approval gates)"]
  F -->|no| D

It also knows when a task **shouldn't** be a loop — a one-shot (use `/goal`) or something where success can't be defined even by a human — and says so, instead of building a token-burner.

The loop it builds for you

A loop is four phases plus a stop rule. The magic is the dotted line — **each run reads what past runs learned