skales

Implements a Reason-Act-Observe loop that chains LLM reasoning with tool execution across 15+ AI providers (OpenAI, Anthropic, Ollama, etc.). The agent maintains a unified provider abstraction layer that normalizes function-calling schemas and response formats, enabling seamless provider switching without code changes. Tool execution results feed back into the reasoning loop for iterative refinement.

Implements an Observe-Orient-Decide-Act state machine that enables fully autonomous task execution with built-in error detection and self-correction. The agent observes task outcomes, re-orients its understanding if results deviate from expectations, decides on corrective actions, and re-executes. Safe Mode requires explicit user approval before autonomous actions modify system state.

Integrates with calendar systems (Google Calendar, Outlook, iCal) and email (IMAP/SMTP) to enable agents to read schedules, propose meetings, send emails, and manage tasks. Planner AI is a specialized agent that understands calendar context and can autonomously schedule meetings, send reminders, and coordinate across attendees. Supports natural language scheduling (e.g., 'schedule a meeting with John next Tuesday at 2 PM').

A persistent desktop mascot (animated character) that represents the agent's state and personality. The Buddy uses a Finite State Machine (FSM) to transition between states (idle, thinking, speaking, error) with corresponding animations and sounds. Notifications are routed through the Buddy (desktop toast, sound, animation) with intelligent prioritization. The Buddy can be clicked to open the chat interface or dismissed.

A specialized code generation and review system that coordinates multiple AI models for different coding tasks. One model generates code, another reviews it for bugs and style, a third optimizes for performance. Supports 40+ programming languages with language-specific linting and formatting. Integrates with local development environments (Git, package managers, test runners) to validate generated code.

Enables multiple Skales instances on a local network to discover each other via mDNS (Bonjour) and coordinate as a swarm. Agents can delegate tasks to peers, share memory and skills, and load-balance work across the network. No central server required — coordination is peer-to-peer. Useful for distributed teams or multi-device setups.

All user data (conversations, memories, API keys, settings, task history) is stored exclusively in ~/.skales-data on the user's machine. No cloud sync, no telemetry, no data transmission to external servers (except to configured LLM providers). Data is organized hierarchically: conversations/, memory/, skills/, tasks/, config/. Users can manually backup or migrate data by copying the directory.

Full internationalization support for UI, agent responses, and system messages across 20+ languages. Locale-specific formatting for dates, times, numbers, and currency. Agent responses can be generated in the user's preferred language. Settings page allows language selection with instant UI refresh.

Runs a background agent process that executes scheduled tasks on cron expressions without requiring the main UI to be active. Tasks are persisted in ~/.skales-data with execution history, enabling long-running automations (e.g., daily reports, periodic monitoring). The agent maintains a task queue that survives app restarts and executes on schedule even if the desktop is locked.

Maintains a dual-layer memory architecture: episodic memory (conversation history with timestamps) and semantic memory (auto-extracted facts, preferences, and user identity). A Memory Scanner runs periodically to extract structured facts from conversations (e.g., 'user prefers Python over JavaScript', 'user works at Acme Corp') and injects them into future agent reasoning. The Soul system maintains a persistent human identity profile that evolves over time.

Integrates a Chromium-based browser controlled by the agent via tool calls (navigate, click, fill forms, extract text). The agent can take screenshots and analyze them with vision-capable models (Claude Vision, GPT-4V) to understand page state and make decisions. Supports JavaScript execution for dynamic content and cookie/session management for authenticated workflows.

Enables multiple specialized agents to collaborate within a single conversation through a Group Chat interface. Each agent has a distinct role (e.g., 'code reviewer', 'documentation writer', 'tester') and can be invoked via @mentions. The Skill Dispatcher routes tasks to agents based on their declared capabilities, and agents can delegate subtasks to each other. Conversation history is shared across the team for context.

Allows users to extend agent capabilities by creating custom skills as .skill.zip packages containing tool definitions, prompts, and optional code. Skill AI is an AI-assisted skill builder that generates skill scaffolding from natural language descriptions. Skills are registered in a local registry and automatically discovered by the agent. Supports both declarative (JSON schema) and imperative (TypeScript) skill definitions.

Implements a full-duplex voice interface with Speech-to-Text (STT), Text-to-Speech (TTS), and Voice Activity Detection (VAD). Users can speak naturally to the agent, and the agent responds with synthesized speech. VAD automatically detects when the user stops speaking and triggers agent response, eliminating the need for manual 'send' buttons. Supports multiple TTS voices and languages.

Grants agents controlled access to the desktop environment: file system operations (read/write/delete with permission checks), keyboard input simulation, mouse movement and clicking, window management, and clipboard access. All file operations are scoped to user-approved directories to prevent accidental data loss. Safe Mode requires approval for destructive operations (delete, overwrite).

Enables remote interaction with the Skales agent through messaging platforms (Telegram, WhatsApp, Discord) and email. Messages are routed to the agent, processed, and responses are sent back through the same channel. Supports media sharing (images, files) and maintains conversation context across platforms. Each platform has a bot bridge that handles authentication and message formatting.