Crowe Logic CLI

A standalone CLI you can customize (UX, commands, workflows) while connecting to your own model endpoints.

Install (dev)

From this folder:

python3 -m venv .venv && source .venv/bin/activate
pip install -e .

Quick Start

Run the interactive configuration wizard:

crowelogic config run

This will guide you through:

• Provider selection (Azure AI Inference, Azure OpenAI, OpenAI-compatible)
• Endpoint and credential setup
• Optional Azure Key Vault integration
• Connection testing

Configuration

You can configure the CLI using environment variables or a config file.

Option A: Config wizard (easiest)

crowelogic config run

Option B: Config file (manual)

Copy the example and fill in your values:

cp .crowelogic.toml.example ~/.crowelogic.toml
# Edit ~/.crowelogic.toml with your endpoint details

The CLI searches for .crowelogic.toml in:

1. Current directory (and parents)
2. Home directory (~/.crowelogic.toml)

Option C: Environment variables

export CROWE_PROVIDER=azure
export CROWE_AZURE_ENDPOINT=https://<your-resource>.openai.azure.com
export CROWE_AZURE_DEPLOYMENT=<deployment-name>
export CROWE_AZURE_API_KEY=<key>
export CROWE_AZURE_API_VERSION=2024-02-15-preview

Commands

Setup and diagnostics

# Interactive setup wizard
crowelogic config run

# Verify connectivity with enhanced diagnostics
crowelogic doctor run

Chat

# Single-shot chat
crowelogic chat run "What is the capital of France?"

# Interactive multi-turn session with streaming and token tracking
crowelogic interactive run

# Save conversation during interactive session
# (Type /save <name> in the REPL)

# Resume a saved conversation
crowelogic history resume my-conversation

Conversation history

# List saved conversations
crowelogic history list

# Load and display a conversation
crowelogic history load my-conversation

# Resume interactive session from history
crowelogic history resume my-conversation

# Delete a conversation
crowelogic history delete my-conversation

Agents and plugins

# List available agents from plugins directory
crowelogic agent list

# Run an agent with a prompt
crowelogic agent run code-reviewer "Review this function"

# List plugins
crowelogic plugins list

# Show plugin details
crowelogic plugins show hookify

Interactive mode commands

In crowelogic interactive run:

• /clear — Clear conversation history
• /save <name> — Save conversation to history
• /system — Show current system prompt
• /exit — Exit (shows session token usage)

Token tracking is automatically displayed after each response and on exit.

For OpenAI-compatible endpoints

If your Azure endpoint uses a generic OpenAI-compatible API (not *.openai.azure.com):

provider = "openai_compatible"

[openai_compatible]
base_url = "https://<your-endpoint>"
model = "<model-id>"
api_key = "<your-api-key>"

CLI Flags Reference

Chat command

crowelogic chat run [OPTIONS] PROMPT

Options:
  -s, --system TEXT   Custom system prompt
  --help              Show help message

Interactive command

crowelogic interactive run [OPTIONS]

Options:
  -s, --system TEXT   Custom system prompt (default: "You are a helpful assistant.")
  --no-stream         Disable streaming responses
  --help              Show help message

Agent command

crowelogic agent run [OPTIONS] AGENT_NAME PROMPT

Options:
  -f, --file PATH     Include file content as context
  --no-stream         Disable streaming responses
  --help              Show help message

Examples:
  crowelogic agent run code-reviewer "Review this code" -f src/main.py
  crowelogic agent run ./my-agent.md "Help me with this task"

History command

crowelogic history resume [OPTIONS] NAME

Options:
  --no-stream         Disable streaming responses
  --help              Show help message

Config command

crowelogic config show      # Display current configuration
crowelogic config path      # Show config file search paths
crowelogic config init      # Create a template config file

Azure Key Vault Integration

Store API keys securely using Azure Key Vault:

[azure]
endpoint = "https://my-resource.openai.azure.com"
deployment = "gpt-4"
api_key = "keyvault://my-vault/my-secret-name"

The CLI uses DefaultAzureCredential for authentication, which supports:

• Environment variables (AZURE_CLIENT_ID, AZURE_TENANT_ID, AZURE_CLIENT_SECRET)
• Managed Identity (when running in Azure)
• Azure CLI (az login)
• Visual Studio Code credentials

Troubleshooting

Configuration errors

"Missing required config for Azure provider"

Ensure all required fields are set. For Azure:

export CROWE_AZURE_ENDPOINT=https://<resource>.openai.azure.com
export CROWE_AZURE_DEPLOYMENT=<deployment-name>
export CROWE_AZURE_API_KEY=<api-key>

Or in .crowelogic.toml:

provider = "azure"

[azure]
endpoint = "https://<resource>.openai.azure.com"
deployment = "<deployment-name>"
api_key = "<api-key>"

Connection errors

401 Unauthorized

• Verify your API key is correct
• Check that the key has access to the specified deployment
• For Key Vault: ensure your Azure identity has Get permission on secrets

404 Not Found

• Check the endpoint URL format
• Verify the deployment name exists in your Azure resource
• Ensure the API version is supported

429 Too Many Requests

• You've hit rate limits; wait and retry
• Consider implementing request throttling

Timeout errors

• Check network connectivity
• Verify the endpoint is accessible from your network
• Try increasing timeout if using a slow connection

Diagnostic commands

# Test configuration and connectivity
crowelogic doctor run

# Show current configuration (masks API keys)
crowelogic config show

# Show where config files are searched
crowelogic config path

Development

Install dev dependencies

pip install -e ".[dev]"

Run tests

pytest                    # Run all tests
pytest -v                 # Verbose output
pytest --cov              # With coverage report
pytest tests/test_config.py  # Run specific test file

Code quality

# Type checking
mypy src/crowe_logic_cli

# Linting
ruff check src/crowe_logic_cli

# Format check
ruff format --check src/crowe_logic_cli

Notes

• This project is intentionally minimal; add commands as your org's workflows solidify.
• You can pair this with the repository's existing plugin content (agents/commands/hooks) as source material, but this CLI is designed to run independently.
• Add .crowelogic.toml to your .gitignore to avoid committing API keys.