Test browser pool data source against Kernel

[IlyaasK]

Summary

• add opt-in live API coverage for the kernel_browser_pool data source
• create one uniquely named durable pool and read it by canonical ID and exact name
• verify normalized durable configuration and null or empty optional references
• prove a no-drift follow-up plan
• verify normal Terraform destroy through a coded not_found API response
• keep SDK deletion as an early-failure cleanup fallback
• add the package to the workflow and documented seven-package acceptance matrix

Scope

Acceptance coverage and release-gate documentation only. This PR does not change provider behavior, schemas, SDK contracts, or runtime/session operations.

The fixture uses KERNEL_ALT_PROJECT_ID when available and otherwise uses KERNEL_PROJECT_ID. At least one must be set. Destruction remains non-forceful.

Tests run

• go test -short -timeout=2m ./...
• go vet ./...
• gofmt -l cmd internal
• terraform fmt -check -recursive examples
• bash scripts/check-docs.sh
• bash scripts/check-markdown-links.sh
• git diff --check

Live acceptance

Passed on commit 3259a54 in workflow run 30660137930.

All seven jobs passed. The new browser-pool data-source job completed in 34 seconds and covered ID lookup, exact-name lookup, explicit project scope, normalized durable state, null references, no drift, non-force destroy, and coded not_found verification.

Intentionally unsupported

• runtime counters or leased/session state
• acquire, release, flush, force recovery, screenshots, logs, or live view
• browser-pool writes through the data source

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit d6ec19c. Configure here.}