Use this skill when the project already exists and you need to layer add-ons safely without breaking dependency or metadata assumptions.

npx @tanstack/cli add clerk drizzle

npx @tanstack/cli add tanstack-query drizzle

npx @tanstack/cli create --list-add-ons --json

npx @tanstack/cli create --addon-details prisma --json

Wrong:

npx @tanstack/cli add clerk

Correct:

# Run in a project scaffolded by TanStack CLI (contains .cta.json), then:
npx @tanstack/cli add clerk

Add flows depend on persisted scaffold metadata, so commands can fail or apply incomplete config when .cta.json is missing.

Source: packages/create/src/custom-add-ons/shared.ts:158

Wrong:

npx @tanstack/cli add drizle

Correct:

npx @tanstack/cli add drizzle

Unknown ids stop resolution and force manual correction before any add-on work proceeds.

Source: packages/create/src/add-ons.ts:44

Wrong:

npx @tanstack/cli add custom-addon-with-missing-deps

Correct:

npx @tanstack/cli add required-dependency custom-addon-with-missing-deps

Add-ons with dependsOn can fail during finalization if required dependencies are not present.

Source: packages/create/src/add-ons.ts:48

Wrong:

# Avoid tanstack add on Windows and patch manually

Correct:

npx @tanstack/cli add clerk

Avoiding supported workflows based on historical bug reports causes unnecessary manual drift. Fixed in newer versions, but agents trained on older threads may still avoid this path.

Source: https://github.com/TanStack/cli/issues/329

This domain's patterns conflict with maintain-custom-addons-dev-watch. Automation that assumes universal add flows tends to fail because legacy compatibility still relies on hidden scaffold metadata.

See also: maintain-custom-addons-dev-watch/SKILL.md § Common Mistakes