fix[bug](graph): add execution_id to base Runtime and restore ctx.execution_id in escalation handler

[Antiarin]

Description

Fixes two problems with escalate_to_coder:

1. NodeContext had no execution_id field, so ctx.execution_id raised AttributeError and crashed the agent
2. Commit 76d4d0d worked around the crash by swapping ctx.execution_id to stream_id, but stream_id is actually ctx.node_id (line 232 of event_loop_node.py), so the escalation event was carrying a node name like "intake" instead of the real execution ID like "exec_abc123".
   No downstream code reads this value today so it wasn't visibly broken, but it's incorrect data sitting in the AgentEvent.

This PR adds execution_id as a proper field on NodeContext, wires it through from the runtime, and restores the original ctx.execution_id in the escalation handler.

Type of Change

• Bug fix (non-breaking change that fixes an issue)

Related Issues

Fixes #5180

Changes Made

• Added execution_id: str = "" field to NodeContext dataclass (node.py)
• Added execution_id property to StreamRuntimeAdapter to expose the private _execution_id (stream_runtime.py)
• Added execution_id property to base Runtime returning "" so both runtime types share the same interface (core.py)
• Wired self.runtime.execution_id through GraphExecutor._build_context into NodeContext (executor.py)
• Restored execution_id=ctx.execution_id in the escalate_to_coder handler, replacing the stream_id workaround (event_loop_node.py)
• Added execution_id to DummyRuntime in test_graph_executor.py to match the contract
• Added 5 tests covering the execution_id plumbing

Testing

• Unit tests pass (708 passed, 0 failed)
• Lint passes (cd core && ruff check .)

Checklist

• My code follows the project's style guidelines
• I have performed a self-review of my code
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes