AutoHedge REPL Commands: Complete Interactive Shell Reference

The AutoHedge REPL recognizes three built-in command types: help (or ?, h) to display usage tips, quit (or exit, q) to terminate the session, and arbitrary text inputs that are interpreted as financial hedging tasks and passed to the core engine.

The AutoHedge repository by The-Swarm-Corporation provides a lightweight command-line interface for automated financial analysis. The interactive REPL is implemented in autohedge/cli.py, offering a deliberately minimal command set that routes user inputs efficiently between the interface and the core hedging engine.

Built-in AutoHedge REPL Commands

The REPL normalizes all input to lower-case and matches it against three specific command categories before processing.

Navigation and Control Commands

The shell recognizes two groups of control commands for user assistance and session management:

• help, ?, h — Displays the REPL tip list, showing available commands and usage hints. This logic resides in lines 64–68 of autohedge/cli.py.

• quit, exit, q — Gracefully terminates the REPL session. The exit handler is implemented in lines 60–62 of autohedge/cli.py.

Task Execution Mode

Any input that does not match the control commands above is treated as a task prompt. The REPL forwards this text directly to AutoHedge.run() for processing. This design routes research queries, stock analysis requests, and hedging calculations through the core engine without requiring specific command prefixes.

REPL Implementation Details

The command parsing logic in autohedge/cli.py follows a straightforward conditional structure after displaying the welcome screen. The implementation converts user input to lower-case to ensure case-insensitive command matching.

When the input matches exit commands, the loop breaks cleanly. Help commands trigger a print statement displaying the tip list. All other inputs fall through to lines 69–78, where the text is passed to the AutoHedge class for execution. This architecture keeps the interface lightweight while allowing the underlying hedging logic in autohedge/main.py to evolve independently.

Practical Usage Examples

Start the REPL

Launch the interactive shell from your terminal:

autohedge

Display Help Information

Request usage tips at the prompt:

> help
  • Enter a task prompt to run (e.g. 'Analyze NVDA for 50k allocation')
  • Type 'quit' or 'exit' to leave
  • Type 'help' or '?' for commands

Execute a Hedging Task

Submit natural language prompts directly to the engine:

> Analyze NVDA for 50k allocation
[Result] … (output from AutoHedge.run())

Terminate the Session

Exit the REPL gracefully:

> quit
Goodbye.

Key Source Files

Understanding the REPL architecture requires familiarity with these components:

• autohedge/cli.py — Implements the REPL loop, command normalization, and the welcome screen interface.

• autohedge/__init__.py — Exposes the AutoHedge class that the REPL instantiates to process task prompts.

• autohedge/main.py — Contains the core AutoHedge.run() method that executes hedging logic for non-command inputs.

• autohedge/prompts.py — Stores default templates and prompt structures leveraged by the REPL and engine.

Summary

• The AutoHedge REPL supports exactly three command types: help variants, exit variants, and free-form task prompts.
• Control commands (help, quit and their aliases) are handled directly in autohedge/cli.py at lines 60–68.
• Task prompts bypass command parsing and route to AutoHedge.run() in autohedge/main.py via lines 69–78 of the CLI module.
• Input is case-insensitive due to lower-case normalization before command matching.
• The minimalist design separates interface concerns from hedging logic, allowing the core engine to evolve without CLI changes.

Frequently Asked Questions

How do I exit the AutoHedge REPL?

Type quit, exit, or the alias q at the prompt. The REPL detects these commands in lines 60–62 of autohedge/cli.py and breaks the input loop gracefully, printing "Goodbye." before returning to the terminal.

What happens if I type a command not recognized by the REPL?

Any text that does not match the help or quit aliases is treated as a task prompt. The REPL passes this input unchanged to AutoHedge.run(), which interprets the text as a financial analysis or hedging request. There are no "unknown command" errors in the AutoHedge REPL.

Where is the REPL command logic implemented?

All REPL functionality resides in autohedge/cli.py. The command dispatch logic spans lines 60–78, with exit handlers at lines 60–62, help display at lines 64–68, and task delegation at lines 69–78.

Can I use the REPL for batch processing or scripts?

No, the REPL is designed exclusively for interactive use. For programmatic access, import the AutoHedge class directly from autohedge/__init__.py and call AutoHedge.run() with your task string, bypassing the CLI interface entirely.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}