#237 data query: stepName and jobName fields always empty in CEL results

Bug

When using swamp data query with CEL predicates, the fields stepName and jobName are listed as available in the CEL schema (shown in error messages), but they are always empty strings in query results.

The actual step and job identifiers are stored in the tags map (e.g., tags.step, tags.workflow, tags.workflowRunId), not in the top-level stepName/jobName fields.

Reproduction

# This returns 0 results even though step-17 data exists:
swamp data query 'stepName == "step-17"' --limit 5 --json

# This works correctly:
swamp data query 'tags.step == "step-17"' --limit 5 --json

# The error message lists both stepName AND tags as valid fields:
swamp data query 'workflow == "x"'
# => Available: ... jobName, ... stepName, ... tags, ...

Expected Behavior

Either:

1. Populate stepName and jobName from tags.step and tags.job (or the workflow context) so queries against them work as expected, OR
2. Remove stepName/jobName from the CEL schema since they are never populated, and document that tags.step/tags.workflow etc. should be used instead

Impact

Agents using the documented CEL field names get empty results. This is especially confusing because the error message for unknown fields explicitly suggests these fields are valid. The workaround (using tags.step etc.) is discoverable but not obvious.

Environment

• swamp version: 20260504.233645.0-sha.430c1535
• PR #1297 added the CEL data query domain layer