Operators
Data Transformer
Productionv1.5.1New
Posted May 13, 2026
Map, reshape and validate structured data between systems using schema-locked operator configs.
Define field mappings, type coercions and validation rules once. The operator enforces them on every run, no ad-hoc transformation logic in prompts.
7
Transform types
6
Formats
50k rows/s
Throughput
100%
Audit events
At a glance
Details
The essentials for Data Transformer: everything published so far.
Category
Operators
Status
Production
Version
v1.5.1
Posted
May 13, 2026
Full details for this skill are publishing soon. Contact us for early access to its actions, run steps, and audit coverage.
Under the hood
Skill manifest
The exact instructions this skill ships with, shown in full. Nothing hidden.
manifest.md
---
id: data-transformer
version: v1.5.1
display_name: Data Transformer
category: operators
category_label: Operators
status: Production
gateway: Secure Connector
tags: ["Data", "Schema", "Validation", "ETL", "JSON Schema"]
---
# Data Transformer
> Map, reshape and validate structured data between systems using schema-locked operator configs.
Define field mappings, type coercions and validation rules once. The operator enforces them on every run, no ad-hoc transformation logic in prompts.
## Stats
- **Transform types:** 7
- **Formats:** 6
- **Throughput:** 50k rows/s
- **Audit events:** 100%
## Supported operations
### Field mapping
Rename, move and restructure fields from source schema to target schema.
### Type coercion
Cast strings to numbers, dates, booleans and enums with error handling.
### Row filter
Drop or flag rows that fail conditional expressions before output.
### Column compute
Derive new fields from expressions, lookups or concatenation rules.
### Schema validate
Assert output conforms to a JSON Schema contract before passing downstream.
### Merge arrays
Combine multiple input arrays with dedup, sort and limit controls.
### Format convert
Serialize output to JSON, CSV, NDJSON, Parquet, XML or XLSX.
## How a call flows
1. Raw input (JSON / CSV / XML)
2. Schema lock check (config validated)
3. Transform engine (map + coerce + filter)
4. Validate output (JSON Schema assert)
5. Audit trail (diff + hash signed)
## Input schema
| Field | Type | Required | Description |
| --- | --- | --- | --- |
| `mappings` | `object[]` | yes | [{source_field, target_field, type, default}] |
| `filters` | `object[]` | no | [{field, operator, value}] row-level filter expressions |
| `computed_fields` | `object[]` | no | [{name, expression}] derived column rules |
| `output_format` | `string` | no | "json" \| "csv" \| "ndjson" \| "parquet" \| "xml" \| "xlsx" |
| `output_schema` | `object` | no | JSON Schema to validate output before passing downstream |
| `on_error` | `string` | no | "skip" \| "flag" \| "fail" (default "fail") |
| `dedup_key` | `string` | no | field name used for deduplication in merge operations |
## Audit and evidence
### Schema lock
Transform config is version-pinned, runtime cannot deviate from the approved mapping definition.
### Row-level diff
Input row count, output row count, dropped rows and error rows all written to audit trail.
### cryptographically signed hash
Input payload hash and output payload hash both stored, proving exactly what was transformed.
### Validation failures
Any JSON Schema assertion failures logged with field path and value before workflow is halted.
## Example use case: Invoice Data Pipeline
1. **Invoice extracted.** Document Extractor returns raw JSON with inconsistent field names from vendor PDFs.
2. **Field mapping.** Data Transformer maps vendor-specific field names to canonical ERP schema.
3. **Type coercion.** Date strings converted to ISO 8601, amounts cast to decimal with currency code.
4. **Validation.** Output asserted against ERP JSON Schema, missing required fields halt the workflow.
5. **Downstream.** Clean, validated payload passed to SAP connector for ledger entry creation.
---
*Generated from `src/app/data/skills.ts` (slug: `data-transformer`). Do not hand-edit. Re-run `npm run generate:skill-manifests` after changing the source data.*
Feedback
Reviews
0.0(No ratings yet)
Reviews from teams using Data Transformer.
No reviews yet
Be the first to share how Data Transformer works for your team.
Discussion(0)
No comments yet. Be the first.