MUST use when writing Bun Native scripts. The script must start with //native to run on the native worker.
CLI Commands
Place scripts in a folder.
After writing, tell the user which command fits what they want to do:
wmill script preview <script_path> — default when iterating on a local script. Runs the local file without deploying.
wmill script run <path> — runs the script already deployed in the workspace. Use only when the user explicitly wants to test the deployed version, not local edits.
wmill generate-metadata — regenerate the local .script.yaml (input schema) and .lock (resolved dependencies) for scripts you changed, and refresh their content hashes in wmill-lock.yaml. Local files only — not a deploy. See "Keep metadata in sync" below.
Deploy local changes to the workspace — via git push or wmill sync push depending on how the repo is wired (see the Deploying section in AGENTS.wmill.md). Only suggest/run a deploy when the user explicitly asks to deploy/publish/push — not when they say "run", "try", or "test".
Preview vs run — choose by intent, not habit
If the user says "run the script", "try it", "test it", "does it work" while there are local edits to the script file, use script preview. Do NOT push the script to then script run it — pushing is a deploy, and deploying just to test overwrites the workspace version with untested changes.
Only use script run when:
The user explicitly says "run the deployed version" / "run what's on the server".
There is no local script being edited (you're just invoking an existing script).
Only use sync push when:
The user explicitly asks to deploy, publish, push, or ship.
The preview has already validated the change and the user wants it in the workspace.
Keep metadata in sync after editing
wmill-lock.yaml tracks a content hash for each item. Editing a script's content — most importantly adding or removing an import or changing main's arguments — invalidates that hash and leaves the .lock, the .script.yaml input schema, and the hash row out of date. Run wmill generate-metadata (scoped to what you touched) after such edits so the resolved lock, the auto-generated args UI (driven by .script.yaml), and wmill-lock.yaml all match the code. Leaving them stale produces spurious diffs in git-sync and CI.
This only writes local files (it is not a deploy), but it re-resolves dependencies, so it can bump unpinned versions (the same as deploying from the UI; expected, not a bug). So by default offer it and run it once the user agrees, rather than running it silently after every edit — unless the project's AGENTS.md opts into running metadata automatically (see the "Keeping metadata in sync" preference there). Either way YOU run the command, not the user. After running it, diff the regenerated .lock / .script.lock files and tell the user which dependency versions changed (e.g. requests 2.31.0 → 2.32.0), so they can catch an unwanted bump before deploying — even under Metadata: auto, since it's information, not a confirmation gate. Pin versions in code to keep them fixed.
With no path argument, generate-metadata regenerates only the items whose content hash drifted — not everything. Imports propagate: editing a script that others import marks every importer stale too, so a one-line change to a shared module can regenerate many locks (by design — their locks must reflect the imported code). If it touches more than you expect, run wmill generate-metadata --dry-run — it lists each stale item with a reason (content changed or depends on <path>) without changing anything — then narrow with a path argument (wmill generate-metadata f/foo) or --strict-folder-boundaries.
If the on-disk .lock and .script.yaml are already correct and only wmill-lock.yaml needs its hashes refreshed (hash drift, or bootstrapping missing entries), use wmill generate-metadata rehash — it re-records hashes from disk with no backend round-trip and no dependency changes.
After writing — offer to test, don't wait passively
If the user hasn't already told you to run/test/preview the script, offer it as a one-sentence next step (e.g. "Want me to run wmill script preview with sample args?"). Do not present a multi-option menu.
If the user already asked to test/run/try the script in their original request, skip the offer and just execute wmill script preview <path> -d '<args>' directly — pick plausible args from the script's declared parameters. The shape varies by language: main(...) for code languages, the SQL dialect's own placeholder syntax ($1 for PostgreSQL, ? for MySQL/Snowflake, @P1 for MSSQL, @name for BigQuery, etc.), positional $1, $2, … for Bash, param(...) for PowerShell.
wmill script preview does not deploy, but it still executes script code and may cause side effects; run it yourself when the user asked to test/preview (or after confirming that execution is intended). wmill generate-metadata does not deploy either — it only writes local files (locks, schemas, hashes) — but offer it before running (or run automatically if the project's AGENTS.md opts in), per "Keep metadata in sync" above. Deploying to the workspace (git push or wmill sync push depending on how the repo is wired — see the Deploying section) is the only step that mutates remote state — do it only when the user explicitly asks to deploy/publish/push.
For a visual open-the-script-in-the-dev-page preview (rather than script preview's run-and-print-result), use the preview skill.
Use wmill resource-type list --schema to discover available resource types.
TypeScript (Bun Native)
Native TypeScript execution. Native scripts are Bun scripts that run on the native worker — a lightweight V8 isolate that exposes fetch and the JavaScript standard library — and can be heavily parallelized. Every script MUST start with //native on its first line so Windmill routes it to the native worker; without it the exact same script runs on the regular Bun worker. You may import npm packages and other Windmill scripts (e.g. ./helper.ts) — imports are resolved and bundled just like a regular Bun script — as long as everything (your code and its dependencies) relies only on fetch and the standard library. Libraries that need Node/Bun runtime APIs (filesystem, node:* modules, child processes, native addons) will not work on the native worker; use the regular bun language for those.
On Windmill, credentials and configuration are stored in resources and passed as parameters to main.
Use the RT namespace for resource types:
//nativeexportasyncfunctionmain(stripe: RT.Stripe) {
// stripe contains API key and config from the resource
}
Only use resource types if you need them to satisfy the instructions. Always use the RT namespace.
Before using a resource type, check the rt.d.ts file in the project root to see all available resource types and their fields. This file is generated by wmill resource-type generate-namespace.
Imports
The constraint is the runtime, not the import list. You may import npm packages and relative Windmill scripts; they are resolved and bundled exactly like a regular Bun script. But the native worker only provides fetch and the JavaScript standard library, so any imported code must work using only those. Anything requiring Node/Bun built-ins (node:fs, child_process, the Bun API, native modules) belongs in a regular bun script instead. Use the globally available fetch for HTTP:
windmill-client works on the native worker (its calls go over fetch), so use it as the preferred way to talk to Windmill — reading resources/variables/states, running scripts and flows, and the S3 helpers below (loadS3File, loadS3FileStream, writeS3File, S3Object). It handles auth, the workspace, and the base URL for you. Reserve raw fetch for calling external HTTP APIs that aren't Windmill.
The full windmill-client API reference (every exported function and its signature) is included in this skill below — consult it for the exact method instead of hand-rolling a fetch against the Windmill API.
Preprocessor Scripts
For preprocessor scripts, the function should be named preprocessor and receives an event parameter:
Windmill provides built-in support for S3-compatible storage operations. The wmill.S3Object type covers both the s3://storage/key URI form (s3:///key for the workspace default storage) and the { s3, storage? } record form — always use it instead of redefining your own.
//nativeimport * as wmill from"windmill-client";
// Load file content from S3constcontent: Uint8Array = await wmill.loadS3File(s3object);
// Load file as streamconstblob: Blob = await wmill.loadS3FileStream(s3object);
// Write file to S3constresult: wmill.S3Object = await wmill.writeS3File(
s3object, // Target path (or undefined to auto-generate)
fileContent, // string or Blob
s3ResourcePath // Optional: specific S3 resource to use
);
TypeScript SDK (windmill-client)
Import: import * as wmill from 'windmill-client'
To know who is running the script, read the contextual variables rather than calling the API:
process.env.WM_END_USER_EMAIL || process.env.WM_EMAIL. WM_END_USER_EMAIL is the app viewer when
the run was triggered from an app and empty otherwise (both variables are always defined), WM_EMAIL
is the user the job is permissioned as. WM_USERNAME is the matching username.
workerHasInternalServer(): boolean
/**
Initialize the Windmill client with authentication token and base URL
@param token - Authentication token (defaults to WM_TOKEN env variable)
@param baseUrl - API base URL (defaults to BASE_INTERNAL_URL or BASE_URL env variable)
*/
setClient(token?: string, baseUrl?: string): void
@returns Job result when completed
*/
async waitJob(jobId: string, verbose: boolean = false): Promise
/**
Get the result of a completed job
@param jobId - ID of the completed job
@returns Job result
*/
async getResult(jobId: string): Promise
/**
Get the result of a job if completed, or its current status
@param jobId - ID of the job
@returns Object with started, completed, success, and result properties
*/
async getResultMaybe(jobId: string): Promise
/**
Cancel a queued or running job by ID.
@param jobId - UUID of the job to cancel
@param reason - Optional reason for cancellation
@returns Response message from the cancel endpoint
*/
async cancelJob(jobId: string, reason: string | undefined = undefined): Promise
/**
Run a script asynchronously by its path
@param path - Script path in Windmill
@param args - Arguments to pass to the script
@param scheduledInSeconds - Schedule execution for a future time (in seconds)
@param tag - Override the worker tag the job runs on
@returns Job ID of the created job
*/
async runScriptByPathAsync(path: string, args: Record<string, any> | null = null, scheduledInSeconds: number | null = null, tag: string | null = null): Promise
/**
Run a script asynchronously by its hash
@param hash_ - Script hash in Windmill
@param args - Arguments to pass to the script
@param scheduledInSeconds - Schedule execution for a future time (in seconds)
@param tag - Override the worker tag the job runs on
@returns Job ID of the created job
*/
async runScriptByHashAsync(hash_: string, args: Record<string, any> | null = null, scheduledInSeconds: number | null = null, tag: string | null = null): Promise
/**
Run a flow asynchronously by its path
@param path - Flow path in Windmill
@param args - Arguments to pass to the flow
@param scheduledInSeconds - Schedule execution for a future time (in seconds)
@param doNotTrackInParent - If false, tracks state in parent job (only use when fully awaiting the job)
@param tag - Override the worker tag the job runs on
@returns Job ID of the created job
*/
async runFlowAsync(path: string | null, args: Record<string, any> | null, scheduledInSeconds: number | null = null, // can only be set to false if this the job will be fully await and not concurrent with any other job // as otherwise the child flow and its own child will store their state in the parent job which will // lead to incorrectness and failures doNotTrackInParent: boolean = true, tag: string | null = null): Promise
/**
Resolve a resource value in case the default value was picked because the input payload was undefined
@param obj resource value or path of the resource under the format $res:path
@returns resource value
*/
async resolveDefaultResource(obj: any): Promise
/**
Get the state file path from environment variables
@returns State path string
*/
getStatePath(): string
/**
Set a resource value by path
@param path path of the resource to set, default to state path
@param value new value of the resource to set
@param initializeToTypeIfNotExist if the resource does not exist, initialize it with this type
*/
async setResource(value: any, path?: string, initializeToTypeIfNotExist?: string): Promise
/**
Set the state
@param state state to set
@param path Optional state resource path override. Defaults to getStatePath().
*/
async setState(state: any, path?: string): Promise
/**
Set the progress
Progress cannot go back and limited to 0% to 99% range
@param percent Progress to set in %
@param jobId? Job to set progress for
*/
async setProgress(percent: number, jobId?: any): Promise
/**
Get the progress
@param jobId? Job to get progress from
@returns Optional clamped between 0 and 100 progress value
*/
async getProgress(jobId?: any): Promise<number | null>
/**
Set a flow user state
@param key key of the state
@param value value of the state
*/
async setFlowUserState(key: string, value: any, errorIfNotPossible?: boolean): Promise
/**
Get a flow user state
@param path path of the variable
*/
async getFlowUserState(key: string, errorIfNotPossible?: boolean): Promise
/**
Get the state shared across executions
@param path Optional state resource path override. Defaults to getStatePath().
*/
async getState(path?: string): Promise
/**
Get a variable by path
@param path path of the variable
@returns variable value
*/
async getVariable(path: string): Promise
/**
Set a variable by path, create if not exist
@param path path of the variable
@param value value of the variable
@param isSecretIfNotExist if the variable does not exist, create it as secret or not (default: false)
@param descriptionIfNotExist if the variable does not exist, create it with this description (default: "")
*/
async setVariable(path: string, value: string, isSecretIfNotExist?: boolean, descriptionIfNotExist?: string): Promise
/**
Build a PostgreSQL connection URL from a database resource
@param s3object - S3 object identifying the file to delete (must have s3 set)
@param workspace - Workspace to delete from (defaults to the WM_WORKSPACE env var)
*/
async deleteS3File(s3object: S3Object, workspace: string | undefined = undefined): Promise
/**
Sign S3 objects to be used by anonymous users in public apps
@param s3objects s3 objects to sign
@returns signed s3 objects
*/
async signS3Objects(s3objects: S3Object[]): Promise<S3Object[]>
/**
Sign S3 object to be used by anonymous users in public apps
@param s3object s3 object to sign
@returns signed s3 object
*/
async signS3Object(s3object: S3Object): Promise
/**
Generate a presigned public URL for an array of S3 objects.
If an S3 object is not signed yet, it will be signed first.
@param s3Objects s3 objects to sign
@returns list of signed public URLs
*/
async getPresignedS3PublicUrls(s3Objects: S3Object[], { baseUrl }: { baseUrl?: string } = {}): Promise<string[]>
/**
Generate a presigned public URL for an S3 object. If the S3 object is not signed yet, it will be signed first.
@param s3Object s3 object to sign
@returns signed public URL
*/
async getPresignedS3PublicUrl(s3Objects: S3Object, { baseUrl }: { baseUrl?: string } = {}): Promise
/**
Get URLs needed for resuming a flow after this step
@param approver approver name
@param flowLevel if true, generate resume URLs for the parent flow instead of the specific step.
This allows pre-approvals that can be consumed by any later suspend step in the same flow.
@returns approval page UI URL, resume and cancel API URLs for resuming the flow
*/
async getResumeUrls(approver?: string, flowLevel?: boolean): Promise<{
approvalPage: string;
resume: string;
cancel: string;
}>
/**
Get an OIDC jwt token for auth to external services (e.g: Vault, AWS) (ee only)
@param audience audience of the token
@param expiresIn Optional number of seconds until the token expires
@param {Object} options - The configuration options for the Teams approval request.
@param {string} options.teamName - The Teams team name where the approval request will be sent.
@param {string} options.channelName - The Teams channel name where the approval request will be sent.
@param {string} [options.message] - Optional custom message to include in the Teams approval request.
@param {string} [options.approver] - Optional user ID or name of the approver for the request.
@param {DefaultArgs} [options.defaultArgsJson] - Optional object defining or overriding the default arguments to a form field.
@param {Enums} [options.dynamicEnumsJson] - Optional object overriding the enum default values of an enum form field.
@returns {Promise} Resolves when the Teams approval request is successfully sent.
@throws {Error} If the function is not called within a flow or flow preview.
@throws {Error} If the JobService.getTeamsApprovalPayload call fails.
Usage Example:
await requestInteractiveTeamsApproval({
teamName: "admins-teams",
channelName: "admins-teams-channel",
message: "Please approve this request",
approver: "approver123",
defaultArgsJson: { key1: "value1", key2: 42 },
dynamicEnumsJson: { foo: ["choice1", "choice2"], bar: ["optionA", "optionB"] },
});
Note: This function requires execution within a Windmill flow or flow preview.
*/
async requestInteractiveTeamsApproval({ teamName, channelName, message, approver, defaultArgsJson, dynamicEnumsJson, }: TeamsApprovalOptions): Promise
setWorkflowCtx(ctx: WorkflowCtx | null): void
async sleep(seconds: number): Promise
/**
Execute fn inline and checkpoint the result. On replay the cached value is
returned without re-executing fn.
fn's result is encoded as JSON and decoded back before it is returned, so
the round that runs the body sees the same types every replay sees: a Date
comes back as a string, a Map as {}. {@link Jsonified} is that shape.
*/
async step(name: string, fn: () => T | Promise,): Promise<Jsonified<Awaited>>
/**
Create a task that dispatches to a separate Windmill script.
