feat: add --format json to individual check commands (#54)

[xyaz1313]

Summary

Adds --format json flag to individual check commands (links, orphans, snippets, references, assets). Previously only check all, score, and diff supported JSON output.

Closes #54

Changes

• New helper: _output_json_findings() — consistent JSON serialization for any findings list
• 5 commands updated: check links, check orphans, check snippets, check references, check assets all accept --format json
• Exit codes preserved: JSON mode maintains the same exit-code contract (exit 1 on errors, exit 2 on security breaches, exit 3 on path traversal)
• 6 new E2E tests: Cover JSON output structure and exit code behavior

JSON Output Format

{
  "findings": [{"rel_path": "...", "line_no": 0, "code": "...", "severity": "...", "message": "..."}],
  "summary": {"errors": 0, "warnings": 0, "info": 0, "security_incidents": 0, "security_breaches": 0, "elapsed_seconds": 0.042}
}

Testing

All existing tests pass + 6 new E2E tests.

[PythonWoods-Dev]

Thanks @xyaz1313 — clean contribution! 🛡️

The helper _output_json_findings() is well-designed (DRY, consistent schema across all commands) and the exit code contract is correctly preserved in JSON mode.

One minor formatting fix was needed on test_cli_e2e.py (ruff format), addressed in PR #56. The feature is now documented in zenzic-doc PR #6 and tracked in the [Unreleased] changelog.