How maestro2903/friday works
Hermes Agent is a multi-interface, self-improving AI agent. Users chat through the CLI/TUI or messaging platforms; requests enter a session, run through the agent loop and model, and may invoke tools before returning a response. The README also documents memory and skill learning, scheduled jobs, subagents, remote terminal backends, voice, and research workflows. Those broader capability links are documented in the README; source excerpts are limited, so the diagram avoids implying unverified internal calls.
Main entry points: conversation_loop.py Conversation Loop · base.py Messaging Gateway · tool_executor.py Tool Execution · console_engine.py CLI
Components 20
User Interfaces
- CLI
hermes_cli/console_engine.py - Terminal UI
ui-tui/src/app/turnController.ts - Desktop App
apps/desktop/src/app/desktop-controller.tsx
Agent Runtime
- Session Handling
gateway/session.py - Conversation Loop
agent/conversation_loop.py - Prompt And Context
agent/prompt_builder.py - Model Providers
agent/anthropic_adapter.py
Tools And Delegation
- Tool Execution
agent/tool_executor.py - Terminal Backends
tools - Subagents
agent/agent_runtime_helpers.py - Tool RPC
agent/oneshot.py
Learning And Memory
- Agent Memory
agent/memory_manager.py - Skill Learning
agent/skill_commands.py - Session Search
agent - Trajectory Research
agent/trajectory.py
Channels And Automation
- Messaging Gateway
gateway/platforms/base.py - Scheduled Jobs
hermes_cli/console_engine.py - Voice Features
agent/transcription_provider.py - Messaging Platforms
gateway/platforms
Other components
- User
Connections 22
- User → CLI: starts chat
README.md - User → Terminal UI: starts chat
README.md - User → Desktop App: uses app
README.md - User → Messaging Platforms: sends messages
README.md - CLI → Conversation Loop: starts agent
- Terminal UI → Messaging Gateway: calls gateway
README.md - Desktop App → Messaging Gateway: uses gateway
- Messaging Platforms → Messaging Gateway: routes messages
README.md - Messaging Gateway → Session Handling: maintains sessions
README.md - Session Handling → Conversation Loop: runs turns
- Conversation Loop → Prompt And Context: builds context
- Conversation Loop → Model Providers: requests completion
- Conversation Loop → Tool Execution: dispatches tools
- Tool Execution → Terminal Backends: uses backends
README.md - Tool Execution → Subagents: spawns agents
README.md - Subagents → Tool RPC: calls tools
README.md - Conversation Loop → Agent Memory: uses memory
- Conversation Loop → Skill Learning: curates skills
README.md - Agent Memory → Session Search: recalls sessions
README.md - Messaging Gateway → Scheduled Jobs: delivers scheduled jobs
README.md - Conversation Loop → Voice Features: transcribes voice memos
README.md - Conversation Loop → Trajectory Research: generates trajectories
README.md