Fix system prompts: explicitly instruct model to use ssh_exec tool ...

Without explicit mention in the prompt the model assumed it had to
establish SSH connections itself and refused. Now each tool is listed
with a direct instruction to use it without hesitation.

Filesystem tool: unrestricted mode by default ...

FS_ALLOWED_PATHS=* (default) allows any path, consistent with terminal.
Path check is now lazy (per-request) instead of module-level,
so config is always fresh after restart.
~ expansion added.

Unrestricted terminal mode and SSH tool ...

- Terminal: TERMINAL_ALLOWED_COMMANDS=* (default) runs via shell,
  supports pipes, redirects, subshells; allowlist mode still available
- FS_ALLOWED_PATHS expanded to cover /home /etc /var /opt
- New ssh_exec tool: execute commands on remote hosts via asyncssh,
  supports named connections (ssh_hosts.json) and inline user@host
- ssh_hosts.json gitignored, ssh_hosts.json.example added as reference
- ssh_exec added to server_admin and smart_home profiles

Dark theme and session delete ...

- Switch to dark color palette via CSS custom properties
- Delete button appears on hover for each session item
- Deleting active session resets chat to empty state

Add web chat client ...

- Single-page chat UI served as static files from FastAPI
- Sidebar with profile selector, new chat button, session list
- WebSocket streaming: shows tool calls and text deltas in real time
- Typing indicator, streaming cursor, auto-resize textarea
- Session history loads on switch

Initial implementation of the agent system core ...

- FastAPI server with REST API and WebSocket streaming
- Modular LLM backend abstraction (Ollama implemented, OpenAI stub)
- Tool system: web_search (ddgs), filesystem, http_request, code_exec, terminal
- Agent profiles: smart_home, server_admin, secretary
- Tool-calling loop with concurrent tool execution
- In-memory session store with SessionStore ABC for future persistence
- Registry pattern for tools, profiles, and backends
- Orchestrator stub as foundation for multi-agent scenarios

Initial commit