Add MCP tool import validation script for CI

[CheeYuTan]

Summary

• New .github/scripts/validate_imports.py that validates all MCP tool modules and core library modules can be imported without errors
• Catches broken imports, missing dependencies, and registration issues that would prevent the MCP server from starting
• Validates 17 tool modules + 4 core library modules
• No workspace connection needed — pure import validation

To integrate into CI (requires workflow scope), add to .github/workflows/ci.yml:

  validate-imports:
    name: Validate MCP Imports
    runs-on: linux-ubuntu-latest
    steps:
      - uses: actions/checkout@v6
      - uses: actions/setup-python@v6
        with:
          python-version: "3.11"
      - name: Install packages
        run: |
          pip install -e databricks-tools-core/
          pip install -e databricks-mcp-server/
      - name: Smoke test imports
        run: python .github/scripts/validate_imports.py

Test proof

$ PYTHONPATH=databricks-tools-core:databricks-mcp-server python3 .github/scripts/validate_imports.py

# Core modules that don't need fastmcp imported successfully:
# databricks_tools_core.auth ✓
# databricks_tools_core.agent_bricks ✓
# databricks_tools_core.lakebase_autoscale ✓

# MCP tool modules correctly report missing fastmcp dependency
# (would pass after pip install -e in CI)

Test	Result
Script runs and reports import status	PASS
Core modules without external deps import cleanly	PASS
Missing deps (fastmcp, sqlglot) correctly reported as failures	PASS
Exit code 1 on failures, 0 on success	PASS