test: add payload offloader runners and integration coverage

[zhongkechen]

Stack

Layer	PR	Scope
Architecture	#678	ADR-006 and design decision
1	#649	Core API, envelopes, runtime, operations
2	#681	Filesystem implementation and retries
3	#682 (this PR)	Testing utilities and integration coverage
4	#683	Examples, E2E infrastructure, and implementation docs

Scope

• make local and cloud test runners resolve offloaded execution and operation payloads lazily
• retain per-operation payload policies and history operation subtype/attempt identity
• add payload-aware TestResult and TestOperation inspection
• add integration coverage for root/step/error replay, operation overrides, map/parallel/child behavior, callbacks, chained invokes, caching, retries, and failure boundaries
• stabilize concurrency integration assertions for valid early-completion outcomes

Intentionally excluded:

• examples and EFS infrastructure
• README, configuration, and wire-format documentation

Validation

• full eight-module Maven reactor on Java 17
• git diff --check
• mvn spotless:check

Related to #463.

[github-actions]

Codex AI review

No actionable findings. Residual risk: cloud payload resolution is unit-tested at the history processor layer, without end-to-end cloud coverage.

Reviewed commit 9f353ea2a1d728885c69a203b7733e728ec7bc2e. Workflow run