Command Decision Tree

Use this page to choose the next command without reading the whole CLI reference.

Pick your path

Situation	Run this	Why
You have a raw MCP URL or stdio command	`pluxx init --from-mcp <source> --yes`	Creates the first source project without invoking a host agent
You already configured the MCP in Claude Code, Cursor, Codex, or OpenCode	`pluxx discover-mcp` then `pluxx init --from-installed-mcp <host:name> --yes`	Imports the existing MCP shape without copying literal secret values
You want import, refinement, build, install, and verification in one flow	`pluxx autopilot --from-mcp <source> --runner codex --mode standard --install --install-target codex --trust`	Fastest end-to-end proof in one selected host
You already have a Pluxx project and the MCP changed	`pluxx sync --from-mcp <source>`	Refreshes generated MCP-derived content while preserving custom sections
You already have a host-native plugin	`pluxx migrate <path>`	Converts an existing plugin into a Pluxx source project

Normal manual flow

Use this when you want to inspect each stage:

npx @orchid-labs/pluxx init --from-mcp https://example.com/mcp --yes --name acme
cd acme
npx @orchid-labs/pluxx doctor
npx @orchid-labs/pluxx lint
npx @orchid-labs/pluxx eval
npx @orchid-labs/pluxx test

Add local install proof:

npx @orchid-labs/pluxx test --install
npx @orchid-labs/pluxx verify-install --target codex

If the plugin contains local hook commands, review them first, then pass --trust:

npx @orchid-labs/pluxx test --install --trust

When to use Agent Mode

Use Agent Mode only after the deterministic scaffold is valid but still reads too generic, too lexical, or too tool-shaped.

npx @orchid-labs/pluxx agent prepare \
  --docs https://docs.example.com/overview \
  --website https://example.com

npx @orchid-labs/pluxx agent run taxonomy --runner codex
npx @orchid-labs/pluxx agent run instructions --runner codex

Use review when you want critique before shipping:

npx @orchid-labs/pluxx agent run review --runner codex --no-verify

What each check answers

Command	Question it answers
`doctor`	Is the source project structurally healthy?
`lint`	Will platform-specific rules reject this?
`eval`	Is the scaffold coherent enough to publish or refine?
`test`	Do config, lint, build, and smoke checks pass together?
`verify-install`	Does the host-visible installed bundle actually exist and work?

For every flag and edge case, use CLI Reference.

Start

Build

Learn

Platform Guides

Examples

Reference

Command Decision Tree

Pick your path

Normal manual flow

When to use Agent Mode

What each check answers

​Pick your path

​Normal manual flow

​When to use Agent Mode

​What each check answers

Pick your path

Normal manual flow

When to use Agent Mode

What each check answers