8aa5547af4
* feat: add managed Telegram login flow and API endpoints - Introduced a new managed Telegram login flow, allowing users to link their Telegram accounts via a deep link. - Added `telegram_login_start` and `telegram_login_check` functions to handle the creation of link tokens and status checks for user linking. - Updated the API with new endpoints for managing Telegram login, including detailed response structures for link token creation and status verification. - Enhanced the `.env.example` file to include a configuration option for the Telegram bot username, facilitating easier setup for developers. * refactor: update Telegram integration to use core RPC for login flow - Replaced the managed DM API with core RPC calls for initiating and checking Telegram login status. - Introduced new API methods `telegramLoginStart` and `telegramLoginCheck` to handle link token creation and verification. - Updated the TelegramConfig and MessagingPanel components to utilize the new login flow, enhancing the user experience and simplifying the codebase. - Adjusted tests to reflect changes in the login process and ensure proper functionality. * docs: update TODO list with new user interaction registration task for memory skill * feat: add ChannelSetupModal for configuring channel integrations - Introduced a new reusable modal component, ChannelSetupModal, for configuring channel integrations such as Telegram and Discord. - The modal can be opened from the Skills page or Settings, enhancing user experience by providing a centralized configuration interface. - Updated MessagingPanel and Skills components to integrate the new modal, allowing users to easily manage their channel settings. - Implemented channel-specific configuration components for better modularity and maintainability. * feat: add ChannelSetupModal for configuring channel integrations - Introduced a new reusable modal component, ChannelSetupModal, for configuring channel integrations such as Telegram and Discord. - The modal can be opened from the Skills page or Settings, enhancing user experience by providing a centralized configuration interface. - Updated MessagingPanel and Skills components to integrate the new modal, allowing users to easily manage their channel settings. - Implemented channel-specific configuration components for better modularity and maintainability. * style: update channel components for improved UI consistency - Refactored styles across various channel configuration components, including ChannelCapabilities, ChannelConfigPanel, ChannelFieldInput, ChannelSelector, DiscordConfig, TelegramConfig, and WebChannelConfig. - Enhanced color schemes and layout for better readability and visual appeal, ensuring a cohesive design across the application. - Updated status badge styles in definitions.ts to align with the new design standards. - Improved error message visibility and loading indicators in Channels page for a more user-friendly experience. * refactor: update WebChannelConfig and tests for improved clarity and functionality - Renamed the `definition` prop in WebChannelConfig to `_definition` for clarity. - Updated test cases in DiscordConfig and TelegramConfig to reflect changes in auth mode labels, ensuring accurate rendering of UI elements. - Adjusted the TelegramConfig test to fix the click event on the connect button, enhancing test reliability. - Modified channel definitions to streamline the managed DM auth mode, improving code organization and maintainability. * fix: enhance WebGL error handling in MeshGradient and RotatingTetrahedronCanvas components - Added error handling in the MeshGradient component to gracefully catch WebGL initialization failures, ensuring the app remains functional when the GPU context is unavailable. - Updated the RotatingTetrahedronCanvas component to verify WebGL context availability before initializing the renderer, preventing crashes and improving user experience. - Modified channel definitions to update auth mode labels for clarity and consistency across the application. * style: update RotatingTetrahedronCanvas colors and animation speed - Changed fill and edge colors in the RotatingTetrahedronCanvas component for improved visual appeal. - Adjusted opacity of the fill material to enhance transparency effects. - Introduced a speed multiplier for rotation animations, allowing for more dynamic visual effects. * refactor: improve WebGL error handling and cleanup in MeshGradient component - Removed unused canvas reference and WebGL context probing to streamline the MeshGradient component. - Enhanced error handling during shader compilation to throw an error when WebGL context may be lost, improving robustness. - Wrapped initialization logic in a try-catch block to gracefully handle failures, ensuring the gradient functionality is disabled if initialization fails. * refactor: simplify Telegram login flow by removing unused link token check - Removed the `check_channel_link_status` method from the `BackendOAuthClient`, streamlining the authentication process. - Updated `telegram_login_check` to directly fetch the user profile and determine link status based on the presence of `telegramId`, enhancing clarity and reducing complexity. - Adjusted logging to reflect the new flow and ensure accurate debugging information. * feat: enhance JSON-RPC method invocation with session expiration handling - Added logic to automatically clear stored session on receiving a 401 error from the backend, improving user experience by ensuring users are redirected to the login page when their session expires. - Introduced a helper function to identify session expiration errors, enhancing code clarity and maintainability. - Refactored the `invoke_method` function to incorporate this new error handling mechanism, ensuring robust session management during RPC calls. * style: update app-dotted-canvas background gradient for improved visibility - Changed the radial gradient background color in the app-dotted-canvas from rgba(0, 0, 0, 0.2) to rgba(0, 0, 0, 0.5), enhancing the contrast and overall visual appeal of the component. * feat: implement channel messaging API with end-to-end testing script - Added a new script `test-channel-messaging.sh` for end-to-end testing of message sending to Telegram via the backend API. - Implemented new backend API methods for sending messages, reactions, creating threads, updating threads, and listing threads in channels. - Enhanced the `BackendOAuthClient` with methods to handle channel messaging operations. - Updated controller schemas and handlers to support the new messaging functionalities, ensuring robust interaction with the backend. - Improved logging for better debugging and tracking of channel operations. * feat: add test-channel-receive script for real-time channel message listening - Introduced a new script `test-channel-receive.mjs` that connects to the backend Socket.IO server and listens for incoming channel messages. - Implemented session token retrieval and validation against the backend, ensuring secure message handling. - Added options for debugging and sending test messages, enhancing the script's usability for testing purposes. - Improved logging for better visibility of connection status and incoming messages. * feat: implement inbound channel message handling for real-time responses - Added functionality to handle inbound messages from channels (e.g., Telegram, Discord) by introducing the `handle_channel_inbound_message` function. - Implemented agent inference loop to process messages and send replies back to the originating channel via the REST API. - Enhanced logging for better tracking of message reception and response handling. - Integrated timeout handling to manage long-running requests effectively. * style: fix prettier formatting for channel components Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
4.4 KiB
todo
-
allow skills to be downloaded from the web
-
allow skills to be written as text formatted files like SKILL.md
-
skills need to specific via JSON-rpc the state changes they make to their state and data files in memory
-
skills need to be able to download custom mcp servers from the web
-
integrate the payments flow properly, skip the connect account page and goto the home page
[] - allow for new skills to be coded on their own [] - allow for multiple instances of a skill to be loaded [] - add a local model that can read through the screen and also go through voice using an API like whisper [] - add a screener recorder that goes through the intefaces in the screen and locally summarizes what is happening and brings more assitance to the user [] clean up the core so that we can run it as a binary on a server or as docker
[x] Separate the binary from the tauri codebase [x] Integrate our custom memory engine into core - sanil [x] Integrate our skills registry into core - steve [x] Integrate accessibility service installation [] Add as a step and setting in the UI - cyrus [x] Remove mentions of zeroclaw from the codebaes [x] Integrate local LLM into core [x] Handle process/deamon properly [x] install the linux philosophy of few modules that do their own thing really well sort of.. [x] Remove android / ios support from the codebase. [x] e2e test to check if daemon and sidecar loading works properly [x] Find a better way to structure the cargo files [x] fix all the rust and cargo issues [] Add icon and app name to the various permission settings - mithil [] add self update based on github release. create a update action on the cli - aniketh [x] for each skill show information on how much data has been synced locally and information on how much syncs have happened so far etc.. - mithil/elvin [x] redo the docs once everything is done. [x] remove unwanted feature flags from the rust binary [] fix the config properly - mithil [] Allow for Migrating from OpenClaw - steve done - to be tested [] allow users to choose which version of LLM model they'd like to choose based on their CPU. better ram and gpu means higher parameter model can be used. - mithil [x] in the client side app, make console.log follow a logger style logging where there's a namespace for every logger (like python) - steve [x] - currently we bundle tauri in the openhumany rust core but that shouldn't really have to be there. it can be completely removed. [x] allow skills to be debuggged from the UI (we shuold try to call various tools or see state from the UI itself)
[] improve the prompts so that it avoid Hallucination. So that we can start to focus more on useful things.I asked a question on Notion. Instead of identifying that it is not connected and should install Notion, it gave me suggestions on fake Notion pages.
-
voiceover functionalities [] fix the overlay [] get it to listen to meetings [] get it to actually use the local whisper model
-
screen intelligence
-
ollama [] fix bug where downloads get iterrupted and it keeps restarting over and over again [] fix bug where download progress each download part instead of the whole model (as download happens in parts) [] once a model has been downloaded we can hide the model window from the IU
-
gmail skill [] allow skills to have their oauth setup locally or credentials enterred manually. in which case we will need to ask for oauth creds and setup the webhook urls ourselves. [] allows skills to have an index so that we can setup functionality to have multiple instances of skill for a user (mulitple gmail accounts etc etc...). [] we need to massively improve the skills development and testing environment so that we can get it as close to production really. so todo that we need to somehow be able to run just the skills runtime from the core rust code within the skills repo so that testing becomes super straightforward (might be heavy, but it'll work) [] use encryption to encrypt data back and forth; especially when working with our version of skills [] massively simplify the skills flow and codebase (less is better)
-
webhook functionality test [] create a debug screen to view and test the available webhooks and also monitor their events
-
memory skill [] should index properly all the things (sanil) [] should properly register user interactions
--- e2e tests to write up
- connecting a channel like telegram/discord works properly
- [] add cmake and tauri driver into the build containers so that we can skip