ThalesGroup/fred

AI Summary: The Fred application's backend is encountering `FileNotFoundError` exceptions when attempting to access CSV files containing simulated cluster consumption data (gco2, USD, wh). The issue arises during requests for carbon consumption, cloud cost, and energy consumption data. The task involves identifying why these CSV files are missing from the expected path ('/app/fred/services/cluster_consumption/data/'), and correcting the file paths or data loading mechanism within the `cluster_consumption_csv_service.py` file.

AI Summary: Debug and fix a UI issue in the Fred project where the landing page briefly flashes when switching between conversations or sessions. The issue occurs when navigating between different agent conversations, causing a momentary display of the empty chat page before the correct conversation loads. The fix likely involves optimizing the UI state management and/or the loading process to prevent this visual glitch.

AI Summary: Consolidate the scattered configuration mechanisms (YAML, JSON, environment variables, etc.) across the Fred project's backend, frontend, and knowledge-flow components into a single, centralized, and easily understandable configuration system. This involves identifying all current configuration methods, designing a unified approach, and implementing the changes across the project.

AI Summary: Enhance the Fred multi-agent AI assistant to provide informative error messages to the user when backend services (MCP or SSE servers) are unreachable. This involves modifying the agents to check for server availability and return appropriate messages to the user interface, instead of silently failing.

AI Summary: Integrate HashiCorp Vault into the Fred project to manage Kubernetes secrets (OpenAI API key and kubeconfig) instead of using `kubectl create secret`. This involves installing Vault, securely storing the secrets within Vault, configuring a Vault secret provider for Kubernetes, and modifying the backend pod deployment to retrieve secrets from Vault during its creation.

AI Summary: The task requires fixing a bug in a multi-agent AI assistant application where different users share chat history when Keycloak authentication is enabled. The solution involves ensuring that session storage is properly cleared upon logout and verifying that FastAPI backend routes, particularly those handling chatbot sessions, are correctly secured to prevent unauthorized access to other users' conversations.