Shell

Shell command execution tool.

Installation

main.py
python
from connectonion import Shell

shell = Shell()

Want to customize? Run co copy shell to get an editable copy.

API

run(command)

Execute a shell command, returns output.

main.py
python
result = shell.run("ls -la")
# Returns: "total 12\ndrwxr-xr-x  8 user staff..."

result = shell.run("git status")
# Returns: "On branch main\nnothing to commit..."

result = shell.run("python --version")
# Returns: "Python 3.11.0"

run_in_dir(command, directory)

Execute command in a specific directory.

main.py
python
result = shell.run_in_dir("npm install", "/path/to/project")
# Returns: "added 100 packages..."

result = shell.run_in_dir("pytest", "/path/to/tests")
# Returns: "5 passed in 0.5s"

Use with Agent

main.py
python
from connectonion import Agent, Shell

shell = Shell()
agent = Agent("coder", tools=[shell])

agent.input("list all python files in current directory")
# Agent runs: shell.run("ls *.py")

agent.input("run the tests")
# Agent runs: shell.run("pytest")

agent.input("install requests package")
# Agent runs: shell.run("pip install requests")

Common Use Cases

main.py
python
# Git operations
shell.run("git status")
shell.run("git add .")
shell.run("git commit -m 'update'")

# Package management
shell.run("pip install requests")
shell.run("npm install express")

# Build and test
shell.run("pytest")
shell.run("npm run build")

# System info
shell.run("pwd")
shell.run("whoami")
shell.run("df -h")

Shell

Installation

API

run(command)

run_in_dir(command, directory)

Use with Agent

Common Use Cases

Star us on GitHub