AxoSequencer: fix step-timeout alarm not falling after timeout clears

[PTKu]

Summary

• AxoStepTimedOutMessenger.Activate now resolves _context and refreshes ActiveContextCount := _context.OpenCycleCount() on every cycle the messenger stays non-Idle, restoring the standard AxoMessenger fall semantics — the step-timeout alarm previously stayed stuck active after the sequencer left the timed-out step.
• Added explicit pass-through Run(IAxoObject) override on AxoSequencer as a placeholder site for the per-cycle _msgStepTimedOut.Serve(THIS) wiring (currently commented).
• Bumped GitVersion.yml next-version 0.56.1 → 0.56.2; matching ### 0.56.2 entries appended to both src/core/docs/CHANGELOG.md and the central CHANGELOG.md.

Test plan

• Run the /core/AxoSequencer showcase scenario: trigger a step timeout, confirm _msgStepTimedOut rises with the step description.
• Let the sequencer advance past the timed-out step (or restore the step) and verify the messenger falls within one cycle — no stale "Step timed out" entry remains in the active list.
• Verify the parent's AggregateMessage(1) count returns to the pre-rise value after the alarm falls.
• Smoke-build the core library (apax ib in src/core/ctrl) and run AXOpen.Core.Tests to confirm no regression in the existing sequencer / messenger tests.
• Verify next-version: 0.56.2 in GitVersion.yml matches the latest ### 0.56.2 heading in both CHANGELOGs.

🤖 Generated with Claude Code