Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
34 changes: 34 additions & 0 deletions agent/slack.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -44,7 +44,7 @@

## Connect additional integrations from Slack

After you connect your Slack workspace, you can ask the agent in Slack to set up any third-party integration that it should use as context. The agent uses [Composio](https://docs.composio.dev/toolkits) to support a broad catalog of third-party services, including issue trackers, knowledge bases, CRMs, and developer tools.

Check warning on line 47 in agent/slack.mdx

View check run for this annotation

Mintlify / Mintlify Validation (mintlify) - vale-spellcheck

agent/slack.mdx#L47

Did you really mean 'CRMs'?

<Note>
You must configure GitHub and GitLab through Mintlify, not Composio. Use the [GitHub app](/deploy/github) or [GitLab integration](/deploy/gitlab) to control which repositories the agent can access.
Expand All @@ -68,7 +68,7 @@

### Expired or revoked connections

Mintlify keeps the status of your connected services in sync with Composio. A connection is disconnected when:

Check warning on line 71 in agent/slack.mdx

View check run for this annotation

Mintlify / Mintlify Validation (mintlify) - vale-spellcheck

agent/slack.mdx#L71

In general, use active voice instead of passive voice ('is disconnected').

- The third-party provider expires or revokes the credentials.
- The provider marks the connection as inactive.
Expand All @@ -83,7 +83,7 @@
- Send direct messages to the agent to use it privately.
- Mention `@mintlify` in a channel to use it publicly and collaboratively.
- Attach files and images directly to your messages for the agent to process and include in your docs.
- Continue conversations in threads to iterate on changes or ask follow-up questions. The agent reuses files and images shared earlier in the thread as context for follow-up requests, so you don't have to reupload them.

Check warning on line 86 in agent/slack.mdx

View check run for this annotation

Mintlify / Mintlify Validation (mintlify) - vale-spellcheck

agent/slack.mdx#L86

Did you really mean 'reupload'?
- Share pull request links with the agent to update related documentation.

## Ask questions about your docs
Expand All @@ -95,7 +95,7 @@

Example questions:
- "Which pages mention the API rate limit?"
- "How do I set up authentication?"

Check warning on line 98 in agent/slack.mdx

View check run for this annotation

Mintlify / Mintlify Validation (mintlify) - vale-spellcheck

agent/slack.mdx#L98

Avoid first-person pronouns such as ' I '.
- "How is our navigation structured?"

## Update documentation
Expand All @@ -104,7 +104,7 @@

- **New request**: Send a direct message to the agent or mention `@mintlify` in a channel with instructions on what to update.
- **Existing thread**: Reply in the thread and mention `@mintlify` with instructions on what to update.
- **With attachments**: Upload images, diagrams, code files, or other documents with your message. The agent automatically processes and includes them in your documentation. In a thread, the agent has access to images and files shared in earlier messages, so you can reference them in follow-up requests without reuploading.

Check warning on line 107 in agent/slack.mdx

View check run for this annotation

Mintlify / Mintlify Validation (mintlify) - vale-spellcheck

agent/slack.mdx#L107

Did you really mean 'reuploading'?

The agent reads the context of the request or thread and creates a pull request in your connected repository with the updates.

Expand Down Expand Up @@ -133,6 +133,40 @@
- **Interruptions**: If you send a follow-up message while the agent is still working, it stops the current task and starts on the new one.
- **Completion**: When the agent finishes, it posts a summary with a link to the pull request and a link to open the changes in the web editor.

## Personal memory

When you ask the agent questions about your docs from Slack, it remembers durable facts about you across threads so you don't have to repeat yourself. Memory is personal to your Mintlify user account and scoped to your organization—other teammates have their own separate memory.

The agent saves a memory when you state something stable about yourself, such as:

- **A response preference**: "I prefer plain-English replies without jargon."
- **An identity or role**: "I'm on the support team."

Check warning on line 143 in agent/slack.mdx

View check run for this annotation

Mintlify / Mintlify Validation (mintlify) - vale-spellcheck

agent/slack.mdx#L143

Avoid first-person pronouns such as 'I'm'.
- **A workflow rule**: "Route customer-facing questions to Plain."

On every new question, the agent loads relevant memories into context before answering, so responses reflect what you've told it before.

The agent does not save:

- Ephemeral details tied to the current task or thread.
- Facts about other people or about your organization as a whole. Those belong in your docs.
- Sensitive personal data such as credentials, addresses, or government IDs.

<Note>
Personal memory applies when you ask the agent questions in Slack. It is not used when the agent captures knowledge from an existing thread to draft a pull request.
</Note>

### Example

In one thread, tell the agent a preference:

> @mintlify I'm a sales engineer—when I ask about our APIs, give me example requests I can paste into a customer email.

Check warning on line 162 in agent/slack.mdx

View check run for this annotation

Mintlify / Mintlify Validation (mintlify) - vale-spellcheck

agent/slack.mdx#L162

Avoid first-person pronouns such as 'I'm'.

Check warning on line 162 in agent/slack.mdx

View check run for this annotation

Mintlify / Mintlify Validation (mintlify) - vale-spellcheck

agent/slack.mdx#L162

Avoid first-person pronouns such as ' I '.

Check warning on line 162 in agent/slack.mdx

View check run for this annotation

Mintlify / Mintlify Validation (mintlify) - vale-spellcheck

agent/slack.mdx#L162

Avoid first-person pronouns such as 'me'.

Check warning on line 162 in agent/slack.mdx

View check run for this annotation

Mintlify / Mintlify Validation (mintlify) - vale-spellcheck

agent/slack.mdx#L162

Avoid first-person pronouns such as ' I '.

In a later thread, ask a docs question:

> @mintlify How does pagination work on the search endpoint?

The agent recalls your role and tailors the answer with copy-ready example requests, without you needing to restate the preference.

## Roles and permissions

The agent's capabilities in Slack match users' [roles](/dashboard/roles).
Expand Down
38 changes: 38 additions & 0 deletions es/agent/slack.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -113,6 +113,44 @@ Mientras el agente trabaja en tu solicitud, proporciona retroalimentación en vi
- **Interrupciones**: Si envías un mensaje de seguimiento mientras el agente aún está trabajando, detiene la tarea actual y comienza con la nueva.
- **Finalización**: Cuando el agente termina, publica un resumen con un enlace a la solicitud de extracción y un enlace para abrir los cambios en el editor web.

<div id="personal-memory">
## Memoria personal
</div>

Cuando le haces preguntas al agente sobre tus documentos desde Slack, recuerda datos duraderos sobre ti entre hilos para que no tengas que repetirte. La memoria es personal de tu cuenta de usuario de Mintlify y está acotada a tu organización: otros miembros del equipo tienen su propia memoria separada.

El agente guarda un recuerdo cuando expresas algo estable sobre ti, como:

- **Una preferencia de respuesta**: "Prefiero respuestas en lenguaje sencillo, sin jerga."
- **Una identidad o rol**: "Estoy en el equipo de soporte."
- **Una regla de flujo de trabajo**: "Redirige las preguntas de clientes a Plain."

En cada nueva pregunta, el agente carga los recuerdos relevantes en el contexto antes de responder, para que las respuestas reflejen lo que ya le has contado.

El agente no guarda:

- Detalles efímeros vinculados a la tarea o al hilo actual.
- Datos sobre otras personas o sobre tu organización en su conjunto. Eso pertenece a tus documentos.
- Datos personales sensibles, como credenciales, direcciones o documentos de identidad.

<Note>
La memoria personal se aplica cuando le haces preguntas al agente en Slack. No se utiliza cuando el agente captura conocimiento de un hilo existente para redactar una solicitud de extracción.
</Note>

<div id="example">
### Ejemplo
</div>

En un hilo, dile al agente una preferencia:

> @mintlify Soy ingeniero de ventas: cuando te pregunte sobre nuestras APIs, dame ejemplos de solicitudes que pueda pegar en un correo a un cliente.

En un hilo posterior, haz una pregunta sobre los documentos:

> @mintlify ¿Cómo funciona la paginación en el endpoint de búsqueda?

El agente recuerda tu rol y adapta la respuesta con ejemplos de solicitudes listas para copiar, sin que tengas que volver a indicar la preferencia.

<div id="roles-and-permissions">
## Roles y permisos
</div>
Expand Down
38 changes: 38 additions & 0 deletions fr/agent/slack.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -114,6 +114,44 @@ Pendant que l'agent travaille sur votre demande, il fournit un retour en direct
- **Interruptions** : Si vous envoyez un message de suivi pendant que l'agent travaille encore, il arrête la tâche en cours et commence la nouvelle.
- **Achèvement** : Lorsque l'agent a terminé, il publie un résumé avec un lien vers la pull request (demande de fusion) et un lien pour ouvrir les modifications dans l'éditeur web.

<div id="personal-memory">
## Mémoire personnelle
</div>

Lorsque vous posez des questions à l'agent sur vos documents depuis Slack, il retient des informations durables à votre sujet d'un fil à l'autre afin que vous n'ayez pas à vous répéter. La mémoire est personnelle à votre compte d'utilisateur Mintlify et limitée à votre organisation : les autres membres de l'équipe ont leur propre mémoire distincte.

L'agent enregistre une information lorsque vous indiquez quelque chose de stable à votre sujet, par exemple :

- **Une préférence de réponse** : "Je préfère des réponses en langage clair, sans jargon."
- **Une identité ou un rôle** : "Je fais partie de l'équipe support."
- **Une règle de flux de travail** : "Redirige les questions des clients vers Plain."

À chaque nouvelle question, l'agent charge les informations pertinentes dans le contexte avant de répondre, afin que ses réponses reflètent ce que vous lui avez déjà dit.

L'agent n'enregistre pas :

- Les détails éphémères liés à la tâche ou au fil en cours.
- Les informations concernant d'autres personnes ou votre organisation dans son ensemble. Cela relève de vos documents.
- Les données personnelles sensibles, telles que des identifiants, des adresses ou des pièces d'identité.

<Note>
La mémoire personnelle s'applique lorsque vous posez des questions à l'agent dans Slack. Elle n'est pas utilisée lorsque l'agent capture des connaissances depuis un fil existant pour rédiger une pull request (demande de fusion).
</Note>

<div id="example">
### Exemple
</div>

Dans un fil, indiquez une préférence à l'agent :

> @mintlify Je suis ingénieur commercial — quand je pose des questions sur nos API, donne-moi des exemples de requêtes que je peux coller dans un e-mail destiné à un client.

Dans un fil ultérieur, posez une question sur les documents :

> @mintlify Comment fonctionne la pagination sur l'endpoint de recherche ?

L'agent se souvient de votre rôle et adapte sa réponse avec des exemples de requêtes prêts à copier, sans que vous ayez besoin de réindiquer votre préférence.

<div id="roles-and-permissions">
## Rôles et autorisations
</div>
Expand Down
Loading