fix(supervisor): keep instance id stable across reconnects

[elezar]

Summary

Keep SupervisorHello.instance_id stable across reconnect attempts made by one supervisor process. The existing protobuf defines the field as a boot/process epoch, so each new process still receives a fresh UUID while connection retries reuse it.

Related Issue

Extracted from #1868 for independent review.

Changes

• Generate the supervisor instance UUID once per run_session_loop.
• Pass the same instance ID into every run_single_session attempt.
• Keep the SupervisorHello wire format unchanged.
• Add focused regression coverage for caller-owned instance-ID reuse.

Testing

• mise run pre-commit passes
  • Attempted inside the Nix environment with reduced mise concurrency. The five-minute run timed out during concurrent workspace Rust builds; remaining compiler processes were resource-killed with SIGKILL, not failed assertions. Completed format, license, Helm, Markdown, Python, packaging, and install-script checks passed.
• nix develop -c cargo test -p openshell-supervisor-process (78 passed, 1 ignored)
• nix develop -c cargo clippy -p openshell-supervisor-process --all-targets -- -D warnings
• Focused instance-ID regression test
• E2E tests added/updated (not applicable; wire format and connection behavior are otherwise unchanged)

Checklist

• Follows Conventional Commits
• Commits are signed off (DCO)

[copy-pr-bot]

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

[elezar]

gator-agent

Blocked

Head SHA: 6485a6d65d366e135ca41b11a5c106939a5cd337

Gator is blocked because GitHub reports this PR has merge conflicts (mergeable_state=dirty, mergeable=false).

Next action: @elezar, please update the branch against main and resolve the merge conflicts. Gator will continue review once the PR is mergeable again.