Dust Webhook Source logo

Dust Webhook Source

OrganizationPopular
dust-tt
dust-webhook-source

Implement a new built-in webhook source provider in `front`. Use when adding a webhook provider such as Linear, GitHub, or Fathom, including provider research, OAuth prerequisites, provider-specific types and client code, preset registration, UI components, and end-to-end testing.

Overview

Publisherdust-tt
Repositorydust
Skill namedust-webhook-source
Stars
1.5K
Forks
370
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by dust-tt on GitHub. Read the source before you install it.

Installation

Install the Dust Webhook Source AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/dust-tt/dust/tree/main/.claude/skills/dust-webhook-source
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/dust-tt/dust.git /tmp/dust
mkdir -p .claude/skills
cp -r /tmp/dust/.claude/skills/dust-webhook-source .claude/skills/dust-webhook-source
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Dust Webhook Source in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Dust Webhook Source on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Dust Webhook Source is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Front Webhook Sources

Implement built-in webhook providers across two trees:

  • front/lib/triggers/built-in-webhooks/<provider>/ — UI-safe: presets, schemas, types, React components
  • front/lib/api/triggers/built-in-webhooks/<provider>/ — server-only: service + provider HTTP client

The boundary is load-bearing. Anything reached from the UI must never transitively import @app/lib/api/config, OAuthAPI, or other backend modules — putting server code under lib/api/ is how we enforce that. Do not re-introduce a webhookService field on the preset or instantiate the service from preset.ts: the leak that motivated the split will come right back.

The work spans provider research, OAuth prerequisites, typed remote metadata, a provider client, the remote webhook service, preset registration, services-map registration, and the setup/details UI.

Hard Prerequisites

Do not start implementation until both of these already exist:

  • front/lib/api/oauth/providers/<provider>.ts
  • core/src/oauth/providers/<provider>.rs

The provider must support the webhooks use case and have scopes that can manage webhooks.

If the provider only allows webhooks to be created manually in its UI, stop. It is not a fit for a built-in webhook source.

Research First

Before writing code, document these answers from the provider docs:

  • how to create a webhook programmatically
  • required auth method and OAuth scopes
  • request and response payloads for create and delete
  • available event types and sample payloads
  • how the incoming event type is identified: header or body field
  • what user-configurable metadata is required, such as team or repository selection
  • whether webhook signatures exist and how to verify them

This research determines the metadata types, client methods, event schemas, and UI.

Files To Touch

Minimal implementation:

  • front/types/triggers/webhooks.ts
  • front/types/triggers/webhooks_client_side.ts
  • front/lib/api/triggers/built-in-webhooks/services.ts (register the service instance)
  • front/lib/api/triggers/built-in-webhooks/<provider>/<provider>_client.ts (server-only)
  • front/lib/api/triggers/built-in-webhooks/<provider>/service.ts (server-only)
  • front/lib/triggers/built-in-webhooks/<provider>/types.ts
  • front/lib/triggers/built-in-webhooks/<provider>/preset.ts
  • front/lib/triggers/built-in-webhooks/<provider>/client_side.ts
  • front/lib/triggers/built-in-webhooks/<provider>/schemas/*
  • front/lib/triggers/built-in-webhooks/<provider>/components/*

Use existing providers as templates:

  • fathom for the simplest shape
  • linear for team-scoped webhook creation
  • github for repository-driven configuration

Workflow

1. Register the provider

In front/types/triggers/webhooks.ts:

  • add the provider to WEBHOOK_PROVIDERS
  • extend WebhookProviderServiceDataMap if the provider needs extra OAuth-fetched data
  • register the preset in WEBHOOK_PRESETS

In front/types/triggers/webhooks_client_side.ts:

  • import the provider client-side preset
  • register it in CLIENT_SIDE_WEBHOOK_PRESETS

In front/lib/api/triggers/built-in-webhooks/services.ts:

  • import the provider service class (from lib/api/..., not lib/...)
  • register an instance in WEBHOOK_SERVICES. This map is the server-only dispatch used by webhook create/delete/service-data endpoints. It must never be imported from UI code.

2. Define metadata and type guards

In types.ts, define:

  • create-time metadata entered by the user
  • persisted metadata that also includes provider webhook ids
  • runtime type guards for both
  • optional additional OAuth data such as teams or repositories

Always keep remote metadata serializable. No functions or class instances.

3. Implement the provider client

<provider>_client.ts lives under front/lib/api/triggers/built-in-webhooks/<provider>/ and should be a thin wrapper over the provider API. It is server-only; do not import it from UI code.

Typical methods:

  • createWebhook(...)
  • deleteWebhook(...)
  • getTeams() / getRepositories() / similar when setup requires selection data

Return typed results and map the provider response into the local metadata shape.

4. Implement the remote webhook service

In front/lib/api/triggers/built-in-webhooks/<provider>/service.ts, implement RemoteWebhookService<"<provider>">. This file is server-only — it will import OAuthAPI and @app/lib/api/config, which must never reach the client bundle.

Core responsibilities:

  • getServiceData(oauthToken) fetches setup data for the form
  • createWebhooks(...) validates metadata, creates provider-side webhooks, and returns updatedRemoteMetadata including webhook ids
  • deleteWebhooks(...) removes provider-side webhooks using the persisted ids

When multiple webhooks may be created, handle partial success explicitly and return user-facing errors without losing the successfully created ids.

5. Define event schemas

Create one schema file per event, or one per coherent event family, under schemas/.

Each event entry should provide:

  • the event name/value exposed in the UI
  • a JSON schema for the payload
  • an optional sample payload when it clarifies the shape

6. Build the presets

Split preset code correctly:

  • preset.ts exports a BaseWebhookPreset — data-only, UI-safe, worker-safe. Do not import React, icons, or the service here. The service is registered separately in WEBHOOK_SERVICES (see step 1).
  • client_side.ts exports a ClientSideWebhookPreset — it spreads the base preset and adds the provider icon plus React components.

Set on the base preset (preset.ts):

  • eventCheck to match how the provider exposes event type
  • the full list of supported events
  • optional webhookPageUrl

Set on the client-side preset (client_side.ts):

  • icon
  • the React components (detailsComponent, createFormComponent, and optionally oauthExtraConfigInput)

7. Build the UI components

Create:

  • CreateWebhook<Provider>Connection.tsx
  • WebhookSource<Provider>Details.tsx

The create form should:

  • rely on front/components/triggers/CreateWebhookSourceWithProviderForm.tsx for the OAuth connection flow via setupOAuthConnection(...)
  • treat connectionId as an input prop; do not recreate the OAuth flow inside the provider-specific form
  • fetch provider-specific setup data with useWebhookServiceData({ owner, connectionId, provider })
  • call onDataToCreateWebhookChange({ connectionId, remoteMetadata })
  • call onReadyToSubmitChange(true) only when configuration is complete

The details component should stay read-only and summarize the persisted remote metadata.

Common Gotchas

  • forgetting to register the data preset in WEBHOOK_PRESETS or the client-side preset in CLIENT_SIDE_WEBHOOK_PRESETS
  • forgetting to register the service in WEBHOOK_SERVICES
  • putting the service or provider client under lib/triggers/... instead of lib/api/triggers/... (leaks OAuthAPI and backend config into the client bundle)
  • importing React, icons, or the service from preset.ts (any of those turn BaseWebhookPreset into a client-only module and reintroduces the bundle leak)
  • putting the icon on the base preset rather than the client-side preset
  • storing non-serializable remote metadata
  • skipping type guards before using remoteMetadata
  • assuming a single webhook when the provider needs one per team or repository
  • ignoring partial success during multi-webhook creation
  • forgetting to make local webhook URLs public during manual testing

Testing

For local testing, expose the local app publicly and set DUST_WEBHOOKS_PUBLIC_URL.

Manual checklist:

  • complete the OAuth flow successfully
  • create the webhook source in the UI without errors
  • verify the webhook exists in the provider dashboard
  • trigger a real provider event and inspect local logs
  • confirm the incoming payload matches the registered schema and event detection logic
  • delete the source and verify the provider-side webhook is removed

References

  • front/lib/resources/webhook_source_resource.ts
  • front/types/triggers/remote_webhook_service.ts
  • front/types/triggers/webhooks_source_preset.ts

Frequently asked questions

What does the Dust Webhook Source AI skill do?

Implement a new built-in webhook source provider in `front`. Use when adding a webhook provider such as Linear, GitHub, or Fathom, including provider research, OAuth prerequisites, provider-specific types and client code, preset registration, UI components, and end-to-end testing.

Why use Dust Webhook Source on TypingMind?

Because you install it once and use it with any model. Dust Webhook Source is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Dust Webhook Source in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/dust-tt/dust/tree/main/.claude/skills/dust-webhook-source. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Dust Webhook Source?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Dust Webhook Source?

As many as you like. As long as a model supports skills, you can use Dust Webhook Source with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Dust Webhook Source AI skill free?

Yes. It is published on GitHub by dust-tt under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇