perf(cli): reuse subtensor connection in issues pending-harvest

[plind-junior]

Why this is necessary: gitt issues pending-harvest opens two WebSocket
connections to the same node for a single command — a bt.Subtensor (used for the
treasury-stake read via the contract client) and then a separate
SubstrateInterface(url=ws_endpoint) for the issue read:

subtensor = bt.Subtensor(network=ws_endpoint)
...
treasury_stake = client.get_treasury_stake()

substrate = SubstrateInterface(url=ws_endpoint)   # second connection to the same node
issues = _read_issues_from_child_storage(substrate, contract_addr, verbose)

The bittensor Subtensor already wraps a SubstrateInterface at
subtensor.substrate, exposing the exact rpc_request / query API the issue
read uses (the contract client reads through subtensor.substrate already). So
the second connection is pure redundancy — extra socket setup/teardown against
the same endpoint for one command.

The fix: read issues through subtensor.substrate and drop the second
SubstrateInterface plus its now-unused import. One command, one connection.
Behavior is unchanged — same data, same code path, just one fewer socket.

Related Issues

N/A

Type of Change

• Bug fix
• New feature
• Refactor
• Documentation
• Other (performance)

Testing

• New tests/cli/test_pending_harvest_single_connection.py: asserts
  pending-harvest --json succeeds, opens no second SubstrateInterface,
  and the issue read reuses subtensor.substrate.
• Existing --json exception-path regression (pending-harvest) still passes.
• uv run pytest — 882 passed.
• uv run pyright / ruff check / vulture on changed files — clean.

Checklist

• Code follows project style guidelines
• Self-review completed
• Changes are documented (if applicable)