Windmill

https://github.com/windmill-labs/windmill-cli-docs
Admin
Windmill is a workflow orchestration platform that enables building, scheduling, and monitoring...
Tokens:164,356
Snippets:1,364
Trust Score:9.4
Update:4 days ago
Show doc for...
Tokens:
Raw
### Install windmill-cli and check version

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/README.md

Installs the Windmill CLI globally via npm and verifies the installation by printing the version. Upgrade later with `wmill upgrade`.

```sh
npm install -g windmill-cli
wmill --version
```

--------------------------------

### Initialize a project directory with wmill init

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/README.md

Creates `wmill.yaml` sync configuration, `AGENTS.md` and `CLAUDE.md` agent prompts, and `.claude/skills/` and `.agents/skills/` per-task guides. Also offers to bind a workspace profile to the current git branch and import git-sync settings from the backend if configured.

```sh
wmill init
```

--------------------------------

### Import libraries

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Libraries are installed automatically, so do not show installation instructions. This example shows typical imports for a Windmill Python script.

```python
import requests
import pandas as pd
from datetime import datetime
```

--------------------------------

### Run, inspect, and debug jobs

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/README.md

Run scripts and flows with inline data or from a file, list failed jobs, and retrieve job details and logs. Logs and flow steps stream as the job runs; for flow failures, `wmill job get` shows the step tree with sub-job IDs for drilling into logs.

```sh
wmill script run u/me/my_script --data '{"foo": "bar"}'
wmill flow   run u/me/my_flow   --data @inputs.json
wmill job list --failed --limit 20
wmill job get  <job_id>
wmill job logs <job_id>
```

--------------------------------

### SQL migration example for creating users table

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/raw-app/SKILL.md

Example migration file sql_to_apply/001_create_users.sql that creates a users table. Use idempotent SQL like CREATE TABLE IF NOT EXISTS; migrations run with no default schema, so tables outside public must be explicitly schema-qualified.

```sql
CREATE TABLE IF NOT EXISTS users (
    id SERIAL PRIMARY KEY,
    email TEXT NOT NULL UNIQUE,
    name TEXT,
    created_at TIMESTAMP DEFAULT NOW()
);
```

--------------------------------

### task usage examples

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-workflow-as-code/SKILL.md

Usage examples for the @task decorator, including basic, with path/timeout/tag, and with retry policy.

```python
@task
async def extract_data(url: str): ...
```

```python
@task(path="f/external_script", timeout=600, tag="gpu")
async def run_external(x: int): ...
```

```python
@task(retry={"attempts": 3, "delay": 30, "multiplier": 2})
async def call_api(payload: dict): ...
```

--------------------------------

### wmill app dev command for apps

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/preview/SKILL.md

Use this command to start the dev server for an app. It runs from the app folder and prints the local app server URL. Long-running, so start it in the background.

```bash
cd <app_path>__raw_app && wmill app dev --no-open --port 4000
```

--------------------------------

### Push individual entities with type-specific commands

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/README.md

Use these commands to push a single entity of a specific type. Each command targets a file path for the corresponding entity type.

```sh
wmill script push   path/to/script.ts
wmill flow   push   path/to/flow.yaml
wmill app    push   path/to/app.yaml
wmill resource push path/to/resource.yaml
```

--------------------------------

### Handling async functions with tokio runtime

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-rust/SKILL.md

Demonstrates how to use async functions (e.g., with tokio) by keeping the `main` function sync and creating the runtime inside it. The example shows fetching a URL with reqwest and returning the response.

```rust
//! ```cargo
//! [dependencies]
//! anyhow = "1.0.86"
//! tokio = { version = "1", features = ["full"] }
//! reqwest = { version = "0.11", features = ["json"] }
//! ```

use anyhow::anyhow;
use serde::Serialize;

#[derive(Serialize, Debug)]
struct Response {
    data: String,
}

fn main(url: String) -> anyhow::Result<Response> {
    let rt = tokio::runtime::Runtime::new()?;
    rt.block_on(async {
        let resp = reqwest::get(&url).await?.text().await?;
        Ok(Response { data: resp })
    })
}
```

--------------------------------

### datatable psql

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Starts a serve listener and launches psql connected to it. Use -n to specify the datatable to connect to.

```APIDOC
## datatable psql

### Description
Starts a serve listener and launches psql connected to it.

### Method
CLI command

### Endpoint
`wmill datatable psql`

### Parameters
#### Options
- `-n --name <name:string>` - Datatable to connect psql to (default: main)
- `--port <port:number>` - Port the proxy listens on (default: first free port in 5433-5500)
- `--host <host:string>` - Bind address for the proxy (default: 127.0.0.1)
- `--password <password:string>` - Password for the temporary Postgres proxy (default: generate a random password at startup)

```

--------------------------------

### Importing packages in Bun scripts

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-bun/SKILL.md

Import npm packages as usual. Libraries are installed automatically.

```typescript
import Stripe from "stripe";
import { someFunction } from "some-package";
```

--------------------------------

### app get <path:string>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Get an app's details by its path. Output can be formatted as JSON for piping to jq.

```APIDOC
## app get <path:string>

### Description
Get an app's details

### Method
CLI command

### Command
`wmill app get <path:string>`

### Parameters
#### Path Parameters
- **path** (string) - Required - The path of the app

### Options
- `--json` - Output as JSON (for piping to jq)

```

--------------------------------

### version()

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Gets the Windmill server version. Returns a version string.

```APIDOC
## version()

### Description
Gets the Windmill server version. Returns a version string.

### Method
Function call

### Returns
- Version string
```

--------------------------------

### launch.json entry for apps

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/preview/SKILL.md

Add this entry to .claude/launch.json to let the MCP tool start the dev server for an app. It runs from the app folder and uses wmill app dev.

```json
{
  "name": "windmill: f/test/my_app",
  "runtimeExecutable": "bash",
  "runtimeArgs": ["-c", "cd f/test/my_app__raw_app && wmill app dev --no-open --port ${PORT:-4001}"],
  "port": 4001,
  "autoPort": true
}
```

--------------------------------

### launch.json entry for flows/scripts

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/preview/SKILL.md

Add this entry to .claude/launch.json to let the MCP tool start the dev server for a flow or script. The entry uses a proxy port and the wmill dev command with --proxy-port.

```json
{
  "name": "windmill: f/test/my_flow",
  "runtimeExecutable": "bash",
  "runtimeArgs": ["-c", "wmill dev --proxy-port ${PORT:-4000} --path f/test/my_flow --no-open"],
  "port": 4000,
  "autoPort": true
}
```

--------------------------------

### denoS3LightClientSettings

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-bun/SKILL.md

Gets S3 client settings from a resource or workspace default. Uses the workspace default if no resource path is provided.

```APIDOC
## denoS3LightClientSettings

### Description
Gets S3 client settings from a resource or workspace default. Uses the workspace default if no resource path is provided.

### Method
async denoS3LightClientSettings(s3_resource_path: string | undefined, workspace: string | undefined = undefined): Promise<DenoS3LightClientSettings>

### Parameters
- **s3_resource_path** (string) - Optional - Path to S3 resource (uses workspace default if undefined)
- **workspace** (string) - Optional - Workspace to read from (defaults to the `WM_WORKSPACE` env var)

### Response
Returns a Promise that resolves to S3 client configuration settings.
```

--------------------------------

### task_flow usage example

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Shows how to create a task that dispatches to a separate Windmill flow. The retry parameter takes the same policy as task. Use inside a workflow function.

```python
pipeline = task_flow("f/etl/pipeline", priority=10)

@workflow
async def main():
    result = await pipeline(input=data)
```

--------------------------------

### Main function structure for Bun scripts

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-bun/SKILL.md

Export a single async function named `main` with typed parameters. Do not call the function; libraries are installed automatically.

```typescript
export async function main(param1: string, param2: number) {
  // Your code here
  return { result: param1, count: param2 };
}
```

--------------------------------

### Python task_script() for existing scripts

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-workflow-as-code/SKILL.md

Creates workflow steps from existing scripts or relative module files. The first example references a local file, the second a script path with a timeout option.

```python
helper = task_script("./helper.py")
existing = task_script("f/data/extract", timeout=600)
value = await helper(input=x)
```

--------------------------------

### Load S3 file with load_s3_file

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Loads a file from the workspace S3 bucket and returns the bytes stream. The example creates an S3Object and decodes the returned bytes to a UTF-8 string.

```python
# s3_obj = S3Object(s3="/path/to/my_file.txt")
# my_obj_content = client.load_s3_file(s3_obj)
# file_content = my_obj_content.decode("utf-8")
# '''
```

--------------------------------

### script show <path:file>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Shows a script's content. Alias for get.

```APIDOC
## script show <path:file>

### Description
Shows a script's content. Alias for get.

### Arguments
- `<path:file>` - Path to the script file
```

--------------------------------

### C# script with NuGet packages using #r directive

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-csharp/SKILL.md

Adds NuGet packages with the #r directive at the top, then uses them in the Main method. The example uses Newtonsoft.Json and RestSharp.

```csharp
#r "nuget: Newtonsoft.Json, 13.0.3"
#r "nuget: RestSharp, 110.2.0"

using Newtonsoft.Json;
using RestSharp;

public class Script
{
    public static object Main(string url)
    {
        var client = new RestClient(url);
        var request = new RestRequest();
        var response = client.Get(request);
        return JsonConvert.DeserializeObject(response.Content);
    }
}
```

--------------------------------

### Start long job and await it (backendAsync + waitJob)

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/raw-app/SKILL.md

Use backendAsync to start a long-running job and get its job id, then waitJob to await the final result. The job id is a string; waitJob rejects if the job failed.

```tsx
import { backendAsync, waitJob } from './wmill';

const jobId = await backendAsync.run_report({ month: '2026-08' }); // a string
const report = await waitJob(jobId);                               // the result itself
```

--------------------------------

### Scaffold new entities with wmill new

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/README.md

Create the correct folder layout and a minimal spec file for scripts, flows, or apps. These commands print next-step hints and pick the right naming conventions for your workspace.

```sh
wmill script new u/me/path --language bun
wmill flow   new u/me/path --summary "..."
wmill app    new u/me/path --summary "..." --framework svelte
```

--------------------------------

### getResultMaybe(jobId: string): Promise<any>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-bun/SKILL.md

Gets the result of a job if completed, or its current status. Returns an object with started, completed, success, and result properties.

```APIDOC
## getResultMaybe(jobId: string): Promise<any>

### Description
Get the result of a job if completed, or its current status.

### Method
Function call

### Parameters
- **jobId** (string) - Required - ID of the job.

### Returns
- **Promise<any>** - Object with started, completed, success, and result properties.

```

--------------------------------

### app dev [app_folder:string]

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Start a development server for building apps with live reload and hot module replacement. Supports various options for port, host, entry point, and more.

```APIDOC
## app dev [app_folder:string]

### Description
Start a development server for building apps with live reload and hot module replacement

### Method
CLI command

### Command
`wmill app dev [app_folder:string]`

### Parameters
#### Path Parameters
- **app_folder** (string) - Optional - The app folder to serve

### Options
- `--port <port:number>` - Port to run the dev server on (will find next available port if occupied)
- `--host <host:string>` - Host to bind the dev server to
- `--entry <entry:string>` - Entry point file (default: index.ts for Svelte, index.tsx otherwise)
- `--no-open` - Don't automatically open the browser
- `--recording` - Frame the app in a shell with a Record button, to capture a replayable session recording of the app under development

```

--------------------------------

### TypeScript taskScript() for existing scripts

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-workflow-as-code/SKILL.md

Creates workflow steps from existing scripts or relative module files. The first example references a local file, the second a script path with a timeout option.

```typescript
const helper = taskScript("./helper.ts");
const existing = taskScript("f/data/extract", { timeout: 600 });
const value = await helper({ input: x });
```

--------------------------------

### Incorrect branchone result references

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-flow/SKILL.md

Do not reference ids of steps that only exist inside a `branchone` branch from outside it. This example shows the incorrect pattern of referencing `results.create_shipment` and `results.create_backorder` directly.

```yaml
expr: results.create_shipment
expr: results.create_backorder
```

--------------------------------

### Correct branchone result reference

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-flow/SKILL.md

Use `results.<branchone_module_id>` to reference the branch result from outside the branch. This example shows a `branchone` step followed by a downstream step that consumes the branch result via a JavaScript input transform.

```yaml
- id: route_order
  value:
    type: branchone
    ...
- id: send_confirmation
  value:
    input_transforms:
      routed:
        type: javascript
        expr: results.route_order
```

--------------------------------

### Load S3 file with load_s3_file_reader

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Loads a file from the workspace S3 bucket and returns a BufferedReader. The example uses a context manager to read the file content.

```python
# '''python
# from wmill import S3Object
# 
# s3_obj = S3Object(s3="/path/to/my_file.txt")
# with wmill.load_s3_file_reader(s3object, s3_resource_path) as file_reader:
#     print(file_reader.read())
# '''
```

--------------------------------

### Java script with Maven dependencies via comments

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-java/SKILL.md

Add Maven dependencies using //requirements: comments at the top of the file. The example declares gson and httpclient dependencies and uses Gson to parse JSON input.

```java
//requirements:
//com.google.code.gson:gson:2.10.1
//org.apache.httpcomponents:httpclient:4.5.14

import com.google.gson.Gson;

public class Main {
    public static Object main(String input) {
        Gson gson = new Gson();
        return gson.fromJson(input, Object.class);
    }
}
```

--------------------------------

### Sync local files with workspace using wmill sync

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/README.md

Use `wmill sync pull` to download workspace changes and `wmill sync push` to upload local changes. Sync is idempotent and diff-aware; `--dry-run` previews changes without applying them. Use `--yaml` (recommended) to keep specs as YAML.

```sh
wmill sync pull     # workspace → local (writes flows, scripts, apps, etc.)
wmill sync push     # local → workspace
```

--------------------------------

### Correct branchall result reference

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-flow/SKILL.md

Use `results.<branchall_module_id>` to reference the branch result from outside the branch. This example shows a `branchall` step with `parallel: true` followed by a downstream step that consumes the branch result via a JavaScript input transform.

```yaml
- id: enrich_parallel
  value:
    type: branchall
    parallel: true
    ...
- id: combine_data
  value:
    input_transforms:
      enrichments:
        type: javascript
        expr: results.enrich_parallel
```

--------------------------------

### task_script usage example

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Shows how to create a task that dispatches to a separate Windmill script. The retry parameter takes the same policy as task. Use inside a workflow function.

```python
extract = task_script("f/data/extract", timeout=600)

@workflow
async def main():
    data = await extract(url="https://...")
```

--------------------------------

### Enable shell completion for wmill (bash, zsh, fish)

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/README.md

Source the completion script for your shell. For bash and zsh, use `source <(wmill completions bash)` or `source <(wmill completions zsh)`. For fish, use `source (wmill completions fish | psub)`. The comment shows the zsh variant inline.

```sh
source <(wmill completions bash)        # bash, zsh: source <(wmill completions zsh)
source (wmill completions fish | psub)  # fish
```

--------------------------------

### Control execution with comment annotations

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-rlang/SKILL.md

Use comment annotations to control execution behavior. Set renv_verbose to show verbose renv output during resolution, renv_install_verbose for package installation output, and sandbox to run in nsjail sandbox (requires nsjail).

```r
#renv_verbose = true        # Show verbose renv output during resolution
#renv_install_verbose = true # Show verbose output during package installation
#sandbox = true              # Run in nsjail sandbox (requires nsjail)
```

--------------------------------

### Python: Approval with step() and wait_for_approval()

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-workflow-as-code/SKILL.md

Name the approval step and generate its URLs inside step() before sending them. get_approval_urls returns URLs bound to that step, the same ones its built-in approve/reject buttons use. Keys must be unique; reusing one raises an error. A minted URL only resumes while its own step is awaiting approval; used at any other moment it is rejected.

```python
urls = await step("urls", lambda: get_approval_urls("manager"))
await step("notify", lambda: send_approval_email(urls["resume"], urls["cancel"]))
approval = await wait_for_approval(key="manager", timeout=3600)
```

--------------------------------

### Correct whileloopflow with stop_after_if on loop module

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-flow/SKILL.md

Use this shape for a whileloopflow when the stop condition can be evaluated on the loop module's result. The loop module's stop_after_if checks result.done after each iteration. The example also shows a counter derived from the iteration index (flow_input.iter.index + 1) and a final step that extracts the last element of the loop result array.

```yaml
- id: loop_until_done
  stop_after_if:
    expr: result.done === true
    skip_if_stopped: false
  value:
    type: whileloopflow
    skip_failures: false
    modules:
      - id: advance_state
        value:
          type: rawscript
          input_transforms:
            count:
              type: javascript
              expr: flow_input.iter.index + 1
- id: return_final_state
  value:
    type: rawscript
    input_transforms:
      final_state:
        type: javascript
        expr: results.loop_until_done[results.loop_until_done.length - 1]
```

--------------------------------

### variable get

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Get a variable's details.

```APIDOC
## variable get <path:string>

### Description
Get a variable's details.

### Method
CLI command

### Command
`variable get <path:string>`

### Parameters
#### Path Parameters
- **path** (string) - Required - The path of the variable.

### Options
- `--json` - Output as JSON (for piping to jq)

```

--------------------------------

### React entrypoint (index.tsx) mounting App into #root

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/raw-app/SKILL.md

The React entrypoint must mount a top-level App into the empty #root div; the preview auto-renders nothing. Keep the UI in App.tsx and use this file only as the mount shim. Never replace it with a bare component or the app renders blank with no error.

```tsx
import React from 'react'
import { createRoot } from 'react-dom/client'
import App from './App'

createRoot(document.getElementById('root')!).render(<App />)
```

--------------------------------

### trigger get

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Get a trigger's details.

```APIDOC
## trigger get <path:string>

### Description
Get a trigger's details.

### Method
CLI command

### Command
`trigger get <path:string>`

### Parameters
#### Path Parameters
- **path** (string) - Required - The path of the trigger.

### Options
- `--json` - Output as JSON (for piping to jq)
- `--kind <kind:string>` - Trigger kind (http, websocket, kafka, nats, postgres, mqtt, amqp, sqs, gcp, azure, email). Recommended for faster lookup

```

--------------------------------

### schedule get <path:string>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/cli-commands.md

Gets a schedule's details.

```APIDOC
## schedule get <path:string>

### Description
Gets a schedule's details.

### Command
`schedule get <path:string>`

### Arguments
- `path` (string) - Required - The schedule path.

### Options
- `--json` - Output as JSON (for piping to jq)

```

--------------------------------

### Python: Inline step with step()

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-workflow-as-code/SKILL.md

Use step() for lightweight inline values that must not change during replay. Use stable, descriptive step names; do not generate step names dynamically.

```python
from datetime import datetime

started_at = await step("started_at", lambda: datetime.now().isoformat())
```

--------------------------------

### resource get <path:string>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/cli-commands.md

Gets a resource's details.

```APIDOC
## resource get <path:string>

### Description
Gets a resource's details.

### Command
`resource get <path:string>`

### Arguments
- `path` (string) - Required - The resource path.

### Options
- `--json` - Output as JSON (for piping to jq)

```

--------------------------------

### @task with path, timeout, and tag

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Configures a task with an external script path, a timeout of 600 seconds, and a 'gpu' tag.

```python
@task(path="f/external_script", timeout=600, tag="gpu")
async def run_external(x: int): ...
```

--------------------------------

### resource-type get <path:string>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/cli-commands.md

Gets a resource type's details.

```APIDOC
## resource-type get <path:string>

### Description
Gets a resource type's details.

### Command
`resource-type get <path:string>`

### Arguments
- `path` (string) - Required - The resource type path.

### Options
- `--json` - Output as JSON (for piping to jq)

```

--------------------------------

### Run wmill app new with flags (non-interactive)

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/raw-app/SKILL.md

Use this command to create a raw app non-interactively. It requires --summary, --path, and --framework; optional flags --datatable and --schema configure data storage. Passing any flag puts the command in non-interactive mode, skipping the datatable wizard and the 'Open in Claude Desktop?' prompt.

```bash
wmill app new \
  --summary "Customer dashboard" \
  --path f/sales/dashboard \
  --framework react19 \
  --datatable main \
  --schema app1
```

--------------------------------

### getWorkspace(): string

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-bun/SKILL.md

Creates a client configuration from environment variables and returns the workspace identifier.

```APIDOC
## getWorkspace(): string

### Description
Create a client configuration from env variables.

### Method
Function call

### Returns
- **string** - The workspace identifier.

```

--------------------------------

### Importing npm packages and Deno standard library

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-deno/SKILL.md

Use `npm:` prefix for npm packages and `https://deno.land/...` for Deno standard library imports.

```typescript
// npm packages use npm: prefix
import Stripe from "npm:stripe";
import { someFunction } from "npm:some-package";

// Deno standard library
import { serve } from "https://deno.land/std/http/server.ts";
```

--------------------------------

### job get

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Get job details. For flows: shows step tree with sub-job IDs.

```APIDOC
## job get

### Description
Get job details. For flows: shows step tree with sub-job IDs.

### Usage
`job get <id:string>`

### Parameters
- `id` (string) - Required - Job ID

### Options
- `--json` - Output as JSON (for piping to jq)
```

--------------------------------

### get(endpoint, raise_for_status=True, **kwargs) -> httpx.Response

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Makes an HTTP GET request to the Windmill API. Returns the HTTP response object.

```APIDOC
## get(endpoint, raise_for_status=True, **kwargs) -> httpx.Response

### Description
Makes an HTTP GET request to the Windmill API.

### Method
GET

### Endpoint
[endpoint]

### Parameters
#### Path Parameters
- **endpoint** (string) - Required - API endpoint path

#### Query Parameters
- **raise_for_status** (boolean) - Optional - Whether to raise an exception on HTTP errors (default: True)
- **kwargs** - Optional - Additional arguments passed to httpx.get

### Response
#### Success Response (200)
- **response** (httpx.Response) - HTTP response object
```

--------------------------------

### setClient(token?: string, baseUrl?: string): void

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-bun/SKILL.md

Initializes the Windmill client with an authentication token and base URL. Defaults to WM_TOKEN and BASE_INTERNAL_URL or BASE_URL environment variables if not provided.

```APIDOC
## setClient(token?: string, baseUrl?: string): void

### Description
Initializes the Windmill client with authentication token and base URL.

### Method
Function call

### Parameters
- **token** (string) - Optional - Authentication token (defaults to WM_TOKEN env variable)
- **baseUrl** (string) - Optional - API base URL (defaults to BASE_INTERNAL_URL or BASE_URL env variable)

### Returns
- **void** - No return value.

```

--------------------------------

### folder get <name:string>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Gets a folder's details by name. Supports JSON output.

```APIDOC
## folder get <name:string>

### Description
Gets a folder's details.

### Arguments
- `<name:string>` - The folder's name

### Options
- `--json` - Output as JSON (for piping to jq)

```

--------------------------------

### flow get <path:string>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Gets a flow's details by its path. Supports JSON output.

```APIDOC
## flow get <path:string>

### Description
Gets a flow's details.

### Arguments
- `<path:string>` - The flow's path

### Options
- `--json` - Output as JSON (for piping to jq)

```

--------------------------------

### Svelte entrypoint (index.ts) mounting App into #root

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/raw-app/SKILL.md

The Svelte entrypoint must mount the App component into the #root div. Keep the UI in App.svelte and use this file only as the mount shim. Never replace it with a bare component or the app renders blank with no error.

```ts
mount(App, { target: document.getElementById('root')! })
```

--------------------------------

### getProgress(jobId)

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-bunnative/SKILL.md

Gets the progress. Optionally specify a job to get progress from. Returns an optional clamped between 0 and 100 progress value.

```APIDOC
## getProgress(jobId)

### Description
Gets the progress. Optionally specify a job to get progress from. Returns an optional clamped between 0 and 100 progress value.

### Method
async getProgress

### Parameters
- **jobId** (any) - Optional - Job to get progress from

### Returns
- **Promise<number | null>** - Optional clamped between 0 and 100 progress value
```

--------------------------------

### Add a workspace profile with wmill workspace add

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/README.md

Walks through adding a workspace profile, a (name, remote URL, workspace id, token) tuple stored under `~/.config/windmill`. Supports multiple profiles; switch with `wmill workspace switch <name>`. Tokens are created in the Windmill UI under User Settings → Tokens; for self-hosted instances, point the remote at your own URL.

```sh
wmill workspace add
```

--------------------------------

### script get <path:file>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Gets a script's details. Supports JSON output for piping to jq.

```APIDOC
## script get <path:file>

### Description
Gets a script's details. Supports JSON output for piping to jq.

### Arguments
- `<path:file>` - Path to the script file

### Options
- `--json` - Output as JSON (for piping to jq)
```

--------------------------------

### get(endpoint, raise_for_status, **kwargs)

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Makes an HTTP GET request to the Windmill API. This method is part of the Windmill Python SDK and can be called directly.

```APIDOC
## get(endpoint, raise_for_status, **kwargs)

### Description
Makes an HTTP GET request to the Windmill API.

### Method
Function call

### Signature
get(endpoint: str, raise_for_status: bool = ..., **kwargs) -> ...

### Parameters
- **endpoint** (str) - Required - API endpoint path
- **raise_for_status** (bool) - Optional - Whether to raise an exception on HTTP errors
- **kwargs** - Optional - Additional arguments passed to httpx.get

### Returns
- **Response** - The HTTP response from the API.
```

--------------------------------

### Configure sandbox and frontend API scopes in raw_app.yaml

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/raw-app/SKILL.md

Set `sandbox: true` to isolate the app frontend from the viewer's Windmill session and declare the frontend API scopes the app needs (e.g., `jobs:run`). A file with a `sandbox` key states both settings; `frontend_api` requires `sandbox: true` or the push fails. A missing `frontend_api` means no scopes, and a file without the `sandbox` key leaves the deployed app's settings unchanged.

```yaml
sandbox: true
frontend_api: [jobs:run]

```

--------------------------------

### audit get <id:string>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Get a specific audit log entry by its ID. Output can be formatted as JSON for piping to jq.

```APIDOC
## audit get <id:string>

### Description
Get a specific audit log entry

### Method
CLI command

### Command
`wmill audit get <id:string>`

### Parameters
#### Path Parameters
- **id** (string) - Required - The ID of the audit log entry

### Options
- `--json` - Output as JSON (for piping to jq)

```

--------------------------------

### backend.<key>(args)

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/raw-app/SKILL.md

Calls a backend runnable and waits for its result. This is the default way to invoke a runnable from the frontend, passing a single object with the runnable's main parameters.

```APIDOC
## backend.<key>(args)

### Description
Calls a backend runnable and waits for its result. This is the default way to invoke a runnable from the frontend, passing a single object with the runnable's main parameters.

### Method
Function call

### Endpoint
`backend.<key>(args)`

### Parameters
- **args** (object) - Required - An object whose keys are the runnable's `main` parameters. If the runnable has no parameters, call with no argument.

### Response
Returns the runnable's result.

### Request Example
```tsx
import { backend } from './wmill';

const user = await backend.get_user({ user_id: '123' });
```
```

--------------------------------

### wait_for_approval: Suspend workflow for external approval

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-workflow-as-code/SKILL.md

Suspends the workflow and waits for an external approval. Pass key to name the step, then get_approval_urls(key) yields the URLs that resume exactly this approval. Without a key the steps are named approval, approval_2, etc. Returns a dict with value, approver, and approved. timeout defaults to 1800 seconds, self_approval defaults to True. skin "minimal" shows approvers only the request instead of the detailed page. Example shows using step to get URLs and send notifications before waiting.

```python
urls = await step("urls", lambda: get_approval_urls("manager"))
await step("notify", lambda: send_email(urls["resume"], urls["cancel"]))
result = await wait_for_approval(key="manager", timeout=3600)
```

```python
async def wait_for_approval(timeout: int = 1800, form: dict | None = None, self_approval: bool = True, key: str | None = None, skin: Literal['detailed', 'minimal'] | None = None, description: str | dict | None = None) -> dict
```

--------------------------------

### waitForApproval(options?: { timeout?: number; form?: object; selfApproval?: boolean; key?: string; skin?: "detailed" | "minimal"; description?: string | object; }): PromiseLike<{ value: any; approver: string; approved: boolean }>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-workflow-as-code/SKILL.md

Suspends the workflow and waits for an external approval. Pass `key` to name the step, then `getApprovalUrls(key)` yields the URLs that resume exactly this approval. `skin: "minimal"` shows approvers only the request instead of the detailed page, and `description` can be a string or a rich value like `{ markdown: "..." }`.

```APIDOC
## waitForApproval(options?: { timeout?: number; form?: object; selfApproval?: boolean; key?: string; skin?: "detailed" | "minimal"; description?: string | object; }): PromiseLike<{ value: any; approver: string; approved: boolean }>

### Description
Suspends the workflow and waits for an external approval. Pass `key` to name the step, then `getApprovalUrls(key)` yields the URLs that resume exactly this approval — route them through your own channel. Without a key the steps are named `approval`, `approval_2`, ... `skin: "minimal"` shows approvers only the request (form and approve/reject) instead of the detailed page with the workflow's details. `description` is shown above the form: a string, or a rich value such as `{ markdown: "..." }`.

### Parameters
#### Options
- **timeout** (number) - Optional - Timeout in seconds.
- **form** (object) - Optional - Form to display to approvers.
- **selfApproval** (boolean) - Optional - Whether self-approval is allowed.
- **key** (string) - Optional - Name of the step.
- **skin** (string) - Optional - "detailed" or "minimal".
- **description** (string | object) - Optional - Description shown above the form.

### Returns
- **PromiseLike<{ value: any; approver: string; approved: boolean }>** - Resolves with the approval result.

### Example
```typescript
const urls = await step("urls", () => getApprovalUrls("manager"));
await step("notify", () => sendEmail(urls.resume, urls.cancel));
const { value, approver } = await waitForApproval({ key: "manager", timeout: 3600 });
```
```

--------------------------------

### state

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Gets the workflow state.

```APIDOC
## state

### Description
Gets the workflow state.

### Method
Function call

### Signature
`def state() -> Any`

### Returns
- **Any** - State value or None if not set
```

--------------------------------

### main function

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-deno/SKILL.md

The main function is the entry point for a Deno script. It must be an async function named 'main' and is called by the Windmill runtime with the script's parameters.

```APIDOC
## main function

### Description
The main function is the entry point for a Deno script. It must be an async function named 'main' and is called by the Windmill runtime with the script's parameters.

### Method
Function

### Signature
`export async function main(param1: string, param2: number)`

### Parameters
- **param1** (string) - Required - First parameter
- **param2** (number) - Required - Second parameter

### Return Value
- **result** (string) - The value of param1
- **count** (number) - The value of param2

### Example
```typescript
export async function main(param1: string, param2: number) {
  // Your code here
  return { result: param1, count: param2 };
}
```
```

--------------------------------

### getFlowUserState

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-bun/SKILL.md

Gets a flow user state.

```APIDOC
## getFlowUserState

### Description
Gets a flow user state.

### Method
async getFlowUserState

### Signature
async getFlowUserState(key: string, errorIfNotPossible?: boolean): Promise<any>

### Parameters
- **key** (string) - Required - Path of the variable
- **errorIfNotPossible** (boolean) - Optional - Whether to error if not possible

### Returns
- **Promise<any>**
```

--------------------------------

### getVariable

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-bun/SKILL.md

Gets a variable by path. Returns the variable value.

```APIDOC
## getVariable

### Description
Gets a variable by path. Returns the variable value.

### Method
async getVariable

### Signature
async getVariable(path: string): Promise<string>

### Parameters
- **path** (string) - Required - Path of the variable

### Returns
- **Promise<string>** - Variable value
```

--------------------------------

### get_boto3_connection_settings(s3_resource_path: str = '') -> Boto3ConnectionSettings

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Returns the settings necessary to initiate an S3 connection using boto3. Without a resource path, the connection goes to the workspace storage through Windmill's S3 proxy when that storage is S3, using the storage's own bucket as Bucket. The proxy serves object reads, writes, deletes, listing and multipart uploads of that bucket only; presigned URLs, other buckets and bucket-level APIs need the storage's own credentials, which passing the resource path returns.

```APIDOC
## get_boto3_connection_settings(s3_resource_path: str = '') -> Boto3ConnectionSettings

### Description
Returns the settings necessary to initiate an S3 connection using boto3. Without a resource path, the connection goes to the workspace storage through Windmill's S3 proxy when that storage is S3, using the storage's own bucket as Bucket. The proxy serves object reads, writes, deletes, listing and multipart uploads of that bucket only; presigned URLs, other buckets and bucket-level APIs need the storage's own credentials, which passing the resource path returns.

### Parameters
#### Parameters
- **s3_resource_path** (str) - Optional - The path to the S3 resource. Defaults to ''.

### Returns
- **Boto3ConnectionSettings** - The connection settings for boto3.
```

--------------------------------

### TypeScript: Approval with step() and waitForApproval()

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-workflow-as-code/SKILL.md

Name the approval step and generate its URLs inside step() before sending them. getApprovalUrls returns URLs bound to that step, the same ones its built-in approve/reject buttons use. Keys must be unique; reusing one raises an error. A minted URL only resumes while its own step is awaiting approval; used at any other moment it is rejected.

```typescript
const urls = await step("urls", () => getApprovalUrls("manager"));
await step("notify", () => sendApprovalEmail(urls.resume, urls.cancel));
const approval = await waitForApproval({ key: "manager", timeout: 3600 });
```

--------------------------------

### get_shared_state

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Gets the state from the shared folder using JSON deserialization.

```APIDOC
## get_shared_state

### Description
Gets the state from the shared folder using JSON deserialization.

### Method
Function call

### Signature
`def get_shared_state(path: str = 'state.json') -> None`

### Parameters
- **path** (str) - Optional - File path for the state (defaults to 'state.json')

### Returns
- **None**
```

--------------------------------

### Specify library dependencies as comments

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-php/SKILL.md

List required libraries as comments before the main function, one per line. Autoloading is already handled.

```php
<?php

// require:
// guzzlehttp/guzzle
// stripe/stripe-php@^10.0

function main() {
    // Libraries are available
}

```

--------------------------------

### wmill dev commands for flows/scripts (direct and proxy modes)

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/preview/SKILL.md

Use these commands when no launch.json-aware runtime is available. Direct mode prints a remote dev-page URL; proxy mode prints a localhost URL that 302s to the remote dev page. Both are long-running, so start them in the background.

```bash
# Direct mode — gives you the remote dev-page URL
wmill dev --path <wmill_path> --no-open

# Proxy mode — gives you a localhost URL that 302s to the remote dev page
wmill dev --proxy-port 4000 --path <wmill_path> --no-open
```

--------------------------------

### get_shared_state_pickle

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Gets the state from the shared folder using pickle deserialization.

```APIDOC
## get_shared_state_pickle

### Description
Gets the state from the shared folder using pickle deserialization.

### Method
Function call

### Signature
`def get_shared_state_pickle(path: str = 'state.pickle') -> Any`

### Parameters
- **path** (str) - Optional - File path for the state (defaults to 'state.pickle')

### Returns
- **Any** - The state value
```

--------------------------------

### Importing windmill-client

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-bun/SKILL.md

Import the Windmill client for platform interactions. Prefer it over raw `fetch` for Windmill operations.

```typescript
import * as wmill from "windmill-client";
```

--------------------------------

### Configure an agent runnable in raw_app.yaml

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/raw-app/SKILL.md

Define an agent runnable in raw_app.yaml to chat with a saved AI agent. The runnable's path is the agent's path. Each call runs one turn of the agent's chat as the app, so a sandboxed app declares no frontend SDK scope and viewers get no consent prompt.

```yaml
runnables:
  chat:
    type: agent
    path: f/support/agent
```

--------------------------------

### getResult(jobId: string): Promise<any>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-bun/SKILL.md

Gets the result of a completed job.

```APIDOC
## getResult(jobId: string): Promise<any>

### Description
Get the result of a completed job.

### Method
Function call

### Parameters
- **jobId** (string) - Required - ID of the completed job.

### Returns
- **Promise<any>** - Job result.

```

--------------------------------

### get_flow_user_state(key: str) -> Any

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Gets the user state of a flow at a given key.

```APIDOC
## get_flow_user_state(key: str) -> Any

### Description
Gets the user state of a flow at a given key.

### Method
Function call

### Parameters
- **key** (str) - Required - The key for the user state

### Returns
- The user state value
```

--------------------------------

### create_token(duration=dt.timedelta(days=1)) -> str

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Creates a new authentication token. Returns the new authentication token string.

```APIDOC
## create_token(duration=dt.timedelta(days=1)) -> str

### Description
Creates a new authentication token.

### Parameters
#### Query Parameters
- **duration** (timedelta) - Optional - Token validity duration (default: 1 day)

### Response
#### Success Response (200)
- **token** (string) - New authentication token string
```

--------------------------------

### getProgress

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-bun/SKILL.md

Gets the progress. Returns an optional clamped between 0 and 100 progress value.

```APIDOC
## getProgress

### Description
Gets the progress. Returns an optional clamped between 0 and 100 progress value.

### Method
async getProgress

### Signature
async getProgress(jobId?: any): Promise<number | null>

### Parameters
- **jobId** (any) - Optional - Job to get progress from

### Returns
- **Promise<number | null>** - Optional clamped between 0 and 100 progress value
```

--------------------------------

### boto3_connection_settings(s3_resource_path)

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Convenient helper that takes an S3 resource as input and returns the settings necessary to initiate an S3 connection using boto3.

```APIDOC
## boto3_connection_settings(s3_resource_path)

### Description
Convenient helper that takes an S3 resource as input and returns the settings necessary to initiate an S3 connection using boto3.

### Method
Function call

### Parameters
- **s3_resource_path** (str) - Optional - S3 resource path

### Returns
Boto3ConnectionSettings
```

--------------------------------

### job result

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Get the result of a completed job (machine-friendly).

```APIDOC
## job result

### Description
Get the result of a completed job (machine-friendly).

### Usage
`job result <id:string>`

### Parameters
- `id` (string) - Required - Job ID
```

--------------------------------

### script bootstrap <path:file> <language:string>

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/cli-commands/SKILL.md

Creates a new script. Alias for new.

```APIDOC
## script bootstrap <path:file> <language:string>

### Description
Creates a new script. Alias for new.

### Arguments
- `<path:file>` - Path to the script file
- `<language:string>` - Language of the script

### Options
- `--summary <summary:string>` - Script summary
- `--description <description:string>` - Script description
```

--------------------------------

### Import Windmill client

Source: https://github.com/windmill-labs/windmill-cli-docs/blob/main/skills/write-script-python3/SKILL.md

Import the windmill client for platform interactions. See the SDK documentation for available methods.

```python
import wmill
```
Tokens:
Raw