SKILL.md

name	testing
description	How to run and extend tests — Mocha unit, sanity API, Jest TypeScript; env and mocks.

Testing skill — @contentstack/management

Commands (from package.json)

Goal	Command
Lint (pre-test hook)	npm run lint
Default npm test	Runs test:api then test:unit per package.json — confirm test:api is defined in your branch (it may be missing); otherwise use test:unit directly
Unit only	npm run test:unit — Mocha test/unit/index.js, nyc coverage
Unit JSON report	npm run test:unit:report:json
Sanity + coverage	npm run test:sanity-test
Sanity, no nyc	npm run test:sanity-nocov
TypeScript / Jest	npm run test:typescript

Always npm run build before sanity if dist/node/contentstack-management.js must match lib/ (sanity imports built output).

Unit tests (test/unit/)

• Mocha + Chai + @babel/register; 30s timeout per script.
• HTTP: mock with Nock, axios-mock-adapter, or stubs — no live CMA.
• Suites are required from test/unit/index.js — add new *-test.js files and register them there.

Sanity tests (test/sanity-check/)

• Live Contentstack CMA calls; orchestrated by test/sanity-check/sanity.js.
• Env: required variables and optional OAuth/Personalize keys are listed at the top of test/sanity-check/utility/testSetup.js (e.g. EMAIL, PASSWORD, HOST, ORGANIZATION).
• Flow: setup creates stack/management token/fixtures where applicable; teardown is handled in the same module — read it before adding destructive tests.

Jest (test/typescript/)

• jest.config.js: ts-jest, coverage to coverage/.
• Use for TypeScript-first checks; keep in sync with public types if asserting SDK shape.

Hygiene

• No committed only / skip on tests meant for CI.
• Never commit real tokens in tests or snapshots.