Build initial official TypeScript SDK

[gerald863]

What changed

• add the official @scrapecreators/sdk TypeScript client
• generate typed request params, response payloads, and client methods for all 186 documented operations
• generate response types from the published OpenAPI sample responses, preferring a formal schema only when one already exists
• keep every sample-derived response property optional and allow undocumented additive fields
• preserve the sample's string type for items[].caption.user.pk
• add API-key auth, typed query/body serialization, structured errors, AbortSignal support, and timeouts that remain active through body reads
• add deterministic OpenAPI sync/code generation, CI, package checks, and a provenance-ready release workflow

Compatibility scope

This SDK does not coerce or runtime-validate API responses. It models the published sample shape while leaving production response behavior unchanged. Third-party payload fields may be added or omitted. We should normalize specific IDs only when the existing public contract requires it rather than coercing unrelated fields globally.

Verification

• ✅ 186 generated SDK methods
• ✅ pnpm check
• ✅ TypeScript compile-time assertions, including items[].caption.user.pk as string | undefined
• ✅ 7/7 runtime tests under Node 22
• ✅ installed-package smoke test under Node 20
• ✅ package tarball generated successfully
• ✅ git diff --check

Publication

This PR does not publish the package. Before the first release, configure npm trusted publishing for:

• package: @scrapecreators/sdk
• repository: ScrapeCreators/scrapecreators-sdk-typescript
• workflow: publish.yml
• environment: npm-publish