cpo-setup

Sets up a repo to run the stack. It fills .cpo-stack/config.yaml with your board facts and runtime sources, writes the tracker adapter, and wires AGENTS.md with a conditional source and task index. Every other skill reads what this one writes for CPO work. Engineering providers such as Pstack also work independently, using their own installation and configuration defaults.

For pstack, follow its README. Cursor installs it with /add-plugin pstack. Use /setup-pstack to choose models and /poteto-mode <task> for engineering. Other hosts need a working installation or adapter for those commands; a Cursor rule alone does not prove that another host loads the model configuration.

Setup and Build resolve commands from the same provider installation. This avoids selecting a different copy when multiple installed skills share a name.

When to use it

• once, before the first use of any other cpo-* skill
• again when you switch trackers or want to restart the config

What it needs

• a git repo with a tracker to point at (Linear or GitHub Issues)
• you, present to answer the setup questions — it confirms real decisions such as tracker, board states, owner, and verify commands

What you get

• a filled .cpo-stack/config.yaml with no TODOs left
• a tracker adapter at docs/agents/issue-tracker.md holding the tracker’s verbs
• an AGENTS.md block that routes agents to the right source and cpo-* skill
• a recorded model-configuration path when the engineering lead uses one

Where it sits

First for CPO work. Every other CPO skill opens by reading the config this skill writes; none of them hardcodes a board fact.

Say this

cpo-setup
set this repo up for the CPO stack
switch the tracker to GitHub Issues