Read and write Greenhouse Recruiting data — candidates, applications, jobs, interviews, scorecards and offers — over the Harvest v3 API.
Greenhouse ships in the w6w first-party pack. It declares 24 actions, 4 health checks, and the host runs its code in a sandbox that never sees the credential.
io.w6w.greenhouseGreenhouse is an applicant tracking system, and this app reads and writes its core recruiting data — candidates, applications, jobs, interviews, scorecards and offers — over Greenhouse’s own Harvest v3 API.
Most of what a workflow needs is one path: find the applications matching a filter, then act on them. This app covers exactly that — creating candidates and their applications, moving an application through pipeline stages, rejecting or hiring it, and filing notes — alongside listing jobs, interviews, scorecards and offers to read the state of a pipeline. Reference lookups for departments, offices, sources, rejection reasons and users let a workflow resolve the real ids Greenhouse’s write endpoints expect rather than guessing labels.
Good for automating candidate intake from a careers page or referral form, moving candidates between pipeline stages based on an external signal such as an assessment result or a scheduling confirmation, syncing interview or offer status into another system, or triggering a rejection notice once a hiring decision is made elsewhere.
Three routes to the same 24 actions. The Workflow tab is generated from Greenhouse's own manifest and carries its real ids, so it is copy-pasteable; the Code and CLI examples are the same call for any action on any app, so every app-specific value in them is a blank you fill in.
create-application Add an application or a prospect record for an existing candidate.
create-candidate Create a candidate, optionally placing them onto a job as their first application.
list-application-stages List the stage history of applications — one row per stage entered, with entered/exited timestamps.
list-applications List applications, scoped by candidate, job, job post, source, stage or lifecycle status.
list-attachments List candidate and application attachments. The returned URLs are signed and short-lived.
list-candidates List candidates (people), optionally filtered by e-mail, tag, privacy, custom field option or timestamp.
list-departments List departments, optionally scoped to one parent or matched by external id.
list-interviews List scheduled interviews, scoped by application, job, organiser or status.
list-job-interview-stages List the interview stages defined on jobs — the ids a move uses.
list-job-posts List job posts, optionally scoped to jobs or boards and filtered by visibility.
list-notes List candidate notes and activity-feed entries, scoped by candidate, application or author.
list-openings List the headcount openings on jobs, filtered by job, state or close reason.
list-rejection-reasons List rejection reasons — the lookup for the id that Reject Application requires.
list-scorecards List interview scorecards, scoped by application, interviewer, submitter or kit.
list-users List Greenhouse users, optionally filtered by office, department, employee id or e-mail.
move-application Move an application to another stage on the same job, or transfer it to a different job.
reject-application Reject an application with a reason, optionally sending or scheduling a rejection e-mail.
update-candidate Patch a candidate's details. Contact-detail and tag fields replace the whole collection.
A workflow step names the app and the action, and the editor fills in the
connection when you pick one. This is the Step shape from the
workflow spec, carrying Greenhouse's real ids.
{
"manifestVersion": "2",
"name": "greenhouse-example",
"steps": [
{
"id": "create-application",
"uses": {
"app": "io.w6w.greenhouse",
"action": "create-application",
"connection": "conn_YOUR_CONNECTION_ID"
},
"with": {
"candidateId": "<candidateId>"
}
}
]
}create-application create-candidate create-note hire-application list-application-stages +19 more actions available
Every app-specific value here is a blank you have to fill in. An
app action is reached through the connection that authenticates it, so the
address is a connection id, not the app id — and connections belong to your account,
so a public page cannot know yours. Create one for Greenhouse, then fill in
the three blanks: conn_YOUR_CONNECTION_ID, the action key, and the
parameters that action declares. The call itself is real — the shape is transcribed
from the studio's own snippet builder, which prints the same kind of blanks — but
nothing in it is specific to Greenhouse. The Workflow tab is where this app's
real ids are.
npm install @w6w/sdkyarn add @w6w/sdkpnpm add @w6w/sdkdeno add npm:@w6w/sdkimport { W6wClient, isActionRun } from "@w6w/sdk";
// Reads W6W_BASE_URL and W6W_TOKEN from the environment when omitted.
const client = new W6wClient();
const envelope = await client.run({
urn: "conn_YOUR_CONNECTION_ID",
action: "create-candidate",
payload: {
firstName: "<value>",
lastName: "<value>",
// preferredName: "<value>",
// company: "<value>",
// title: "<value>",
// emailAddress: "<value>",
// emailType: "<value>",
// phoneNumber: "<value>",
// phoneType: "<value>",
// tags: "<value>",
// jobId: "<value>",
// sourceId: "<value>",
// recruiterId: "<value>",
// coordinatorId: "<value>",
},
});
if (isActionRun(envelope)) console.log(envelope.value); npm install -g @w6w/cli w6w run conn_YOUR_CONNECTION_ID --action create-candidate --payload '{"firstName":"<value>","lastName":"<value>"}' Give an AI agent Greenhouse — without giving it Greenhouse's credentials. One MCP endpoint exposes every app, function and workflow the caller is entitled to, as tools it can discover and run. Access is granted per team while we onboard.
One tool call{
"name": "w6w_invoke",
"arguments": {
"ref": "app:io.w6w.greenhouse#create-application",
"input": {
"candidateId": "<candidateId>"
}
}
}
Every tool names its target with a single ref. The
app: form above doesn't name a connection at all — the
host resolves which of the caller's Greenhouse connections to sign
with, and refuses rather than guesses when the answer is ambiguous.
The token is attached host-side, at the moment of the call. It is never a tool argument, never in the model's context, and never in a transcript — so a prompt injection has nothing to exfiltrate.
Tools are derived per end user from what that person has actually connected and is entitled to — not one shared bot identity carrying the union of everyone's access.
Multi-step work runs on the workflow engine and returns a run handle the agent can poll — retries, branching and state survive the conversation that started them.
Greenhouse's declared health checks are on the surface too, so an agent can tell "the vendor is down" from "your credential expired" before it burns a retry on either.
The MCP surface is part of the hosted platform. Greenhouse itself is MIT, and the runtime that executes it is source-available (FSL).
Greenhouse declares its own checks, so its health is a property of the app rather than something the host guesses at.
Component status from status.greenhouse.io. The verdict comes from the eleven components inside the Greenhouse Harvest API group; Recruiting, Onboarding, the HRIS links, the third-party integrations and the AWS and Fastly infrastructure are reported alongside them but do not set it.
Unauthenticated probe of GET /v3/candidates. A schema-correct 401 is a pass — it proves the route exists and Greenhouse is answering. A 404 means the endpoint has been removed, which is the failure a status page cannot report.
Requests left in the current 30-second window, read from the X-RateLimit-* headers Greenhouse returns on every response.
Greenhouse's status page reports the Harvest API as eleven independent silos, but publishes no way to learn which one an organisation runs on — so the platform check has to roll all eleven together rather than report yours.