🧠 Add Intents and User Confirmations

a. Intent Recognition

Description:
Enable the system to understand user intent from their query, which helps route the request through the correct workflow (e.g., scraping, reporting, KB creation). The goal is to reduce reliance on explicit instructions and improve user experience through intelligent command inference.

Acceptance Criteria:

• Define a set of supported intents (e.g., scrape, report, search, build_kb).
• Extract intent using rule-based NLP or a small classifier model.
• Route workflows based on the detected intent.
• Log the detected intent and provide fallback if confidence is low.

b. User Confirmation Flow

Description:
In cases where the system encounters ambiguous or multi-intent user queries, it should initiate a confirmation dialogue. This interaction will clarify what the user really wants, preventing mistakes and improving trust in automation.

Acceptance Criteria:

• Detect ambiguity or overlapping intents.
• Prompt the user with a clarification question or intent options.
• Proceed only after confirmation is received.
• Handle errors and dead ends gracefully.

📘 Note: Please read and follow the Contribution Guide before starting work on any of the tasks listed below.

Read CONTRIBUTING.md first, then trace the existing FastAPI workflow entry points and how LangGraph and PydanticAI are used. Define the supported intents, routing and confirmation behavior, including low-confidence and dead-end handling; done means the acceptance criteria are implemented and detected intents are logged.