feat(ci): shard integ tests for faster runs.

[Hweinstock]

Problem

Integ tests have become the bottleneck on PR CI runs.
#1487

Solution

• shard them across 4 runners.
• also centralize on same git email that e2e tests use.

Testing

• see tests on PR. No sharding took 11 min, sharding took: 4 min 50 sec. over 50% reduction.

[agentcore-devx-automation]

Claude Security Review: no high-confidence findings. (run)

[github-actions]

Package Tarball

aws-agentcore-0.18.0.tgz

How to install

gh release download pr-1484-tarball --repo aws/agentcore-cli --pattern "*.tgz" --dir /tmp/pr-tarball
npm install -g /tmp/pr-tarball/aws-agentcore-0.18.0.tgz

[agentcore-cli-automation]

Looks good to merge. The sharding approach is correct (vitest --shard=N/M splits at the file level, and 18 integ test files / 4 shards is reasonable), and the new shards are passing on this PR (~2m50s each, vs the previous serial integ run). Email change to ci@amazon.com aligns with e2e-tests.yml and e2e-tests-full.yml as described.

Couple of minor non-blocking observations (no changes needed):

• .github/workflows/pr-tarball.yml:42 still uses the old bedrock-agentcore-npm+ci@amazon.com — outside the stated scope of this PR, but worth a follow-up if the goal is to fully centralize.
• unit-test is sharded 3-ways while integ-test is sharded 4-ways; assuming this reflects relative durations, that's fine.