clara.skills.email_assistant package

Submodules

clara.skills.email_assistant.workflows module

Email Assistant Workflow Module for Clara Core.

Implements the cognitive CRUD loop for email management: 1. Parse Intent (Create, Read, Update, Delete) 2. Extract Entities (Recipients, Subject, Body, Date, Folder, Attachments) 3. Confirm Ambiguity / Guardrails 4. Execute via Gmail Connection tools 5. Format structured markdown response

class clara.skills.email_assistant.workflows.EmailAssistantWorkflow(service=None, access_token=None)[source]

Bases: object

Workflow coordinator for processing natural language email requests.

Orchestrates intent classification, entity extraction, safety confirmation guardrails, tool execution against Gmail APIs, and markdown response generation.

Parameters:

  • service (Any | None)

  • access_token (str | None)

service

Authenticated Gmail API client service.

Type:

Optional[Any]

access_token

OAuth2 access token for user authentication.

Type:

Optional[str]

Example

>>> workflow = EmailAssistantWorkflow(access_token="ya29.a0...")
>>> response = workflow.execute("Draft an email to boss@example.com")
>>> print(response["reply"])
execute(user_prompt, context=None)[source]

Main entry point for executing an email assistant request.

Parameters:

  • user_prompt (str) – Natural language user instruction.

  • context (Optional[Dict[str, Any]]) – Optional conversation context or previous pending action state.

Returns:

Execution result containing:

  • status (str): ‘success’, ‘pending_confirmation’, ‘clarification_needed’, ‘cancelled’, or ‘error’.

  • action (str): Executed action name.

  • reply (str): User-facing markdown formatted reply.

  • data (Dict[str, Any]): Structured output payload.

Return type:

Dict[str, Any]

clara.skills.email_assistant.workflows.handle_email_prompt(user_prompt, access_token=None, service=None, context=None)[source]

Convenience helper to process a natural language prompt with the Email Assistant.

Parameters:

  • user_prompt (str) – The natural language instruction from the user.

  • access_token (Optional[str]) – Optional OAuth2 bearer token.

  • service (Optional[Any]) – Optional pre-authenticated Gmail API client.

  • context (Optional[Dict[str, Any]]) – Optional execution context.

Returns:

Formatted execution results dictionary.

Return type:

Dict[str, Any]

Example

>>> res = handle_email_prompt("Check my unread emails", access_token="ya29...")
>>> print(res["reply"])

Module contents

Email Assistant Skill for Clara Core.

Provides cognitive workflows to parse natural language email commands, draft, send, search, summarize threads, archive, and delete messages using Google Gmail.