Create your own transformer
Build custom transformers for validation, enrichment, or redaction.
The transformer interface
Transformers are functions that receive a context object and return a transformer instance:
The context contains:
The returned instance must implement:
Return values
| Return | Behavior |
|---|---|
{ event } | Continue chain with modified event |
void | Continue chain, event unchanged |
false | Stop chain, event dropped |
{ event, next } | Redirect to a different chain (fan-out) |
{ event, respond } | Continue chain with wrapped respond function |
Minimal example
Push context
The push function receives an event and a push context:
config: Transformer configurationenv: Environment dependencieslogger: Scoped logger for outputid: Transformer identifiercollector: Access to collector instanceingest: Request metadata from source (optional)
Examples
Enrich: add server-side data
Validate: custom check
Using your transformer
Pass-through steps
A path is the multi-step chain through a flow's transformer section. A
pass-through step (short: pass) is a single step within a path that
has no code and no package. The runtime synthesizes its push function for
you, so the step exists purely to wire chain hops, run a cache check, or
apply a declarative mapping.
Three variants exist:
1. Before/next chain only
A named hop in the chain. Useful for reusing a chain prefix across destinations.
"enrichServer": { "before": ["filterBots", "sessionLookup"] }
2. Cache only
A step that runs a cache check, often for dedup or response caching.
"dedup": {
"cache": {
"stop": true,
"rules": [{ "key": ["event.id"], "ttl": 60 }]
}
}
3. Mapping only
A step that runs a declarative event-to-event mapping using Mapping.Config
primitives (policy, per-rule policy, mapping[].name, mapping[].ignore).
"redactPII": {
"mapping": {
"policy": { "user.email": { "value": "[redacted]" } }
}
}
mappingThe mapping field uses the same Mapping.Config shape on destinations and
transformer steps, but the semantic differs by position. On a destination,
mapping shapes the vendor payload. On a transformer step, it mutates the
event itself. Vendor-payload fields (data, per-rule data, silent) are
ignored at the transformer position with a one-time warning at init.
Transformer entries use a closed schema: unknown top-level keys are
errors. This catches typos like { "rules": [], "stop": true } at the top
of a step (forgot the cache: wrapper).
Testing
Package convention
Every walkerOS package includes machine-readable metadata for tooling and discovery.
walkerOS field in package.json
| Field | Required | Description |
|---|---|---|
walkerOS | Yes | Object with type metadata (and platform for sources/destinations) |
Build-time generation
Use buildDev() from the shared tsup config to auto-generate walkerOS.json:
This file contains your package's JSON Schemas and examples, enabling MCP tools and the CLI to validate configurations without installing your package.
Optional: Hints
Packages can export a hints record from src/dev.ts to provide lightweight, actionable context beyond schemas and examples, such as validation behavior, enrichment patterns, or troubleshooting tips. Hints are serialized into walkerOS.json and surfaced via MCP tools. See the walkeros-create-transformer skill for details.
Publishing checklist
-
walkerOSfield in package.json - Keywords include
walkerosandwalkeros-transformer -
buildDev()in tsup.config.ts -
dist/walkerOS.jsongenerated on build -
npm run testpasses -
npm run lintpasses
Next steps
- Learn about creating sources
- Learn about creating destinations