Python SDK - Alshival.Ai

Configure

import alshival

alshival.configure(
    username='your_username',
    api_key='your_api_key',
    resource='https://alshival.dev/u/resource_owner_username/resources/resource_uuid/',  # paste Resource URL from DevTools Resource Details
    cloud_level='ALERT',
    debug=False,
)

Set resource to the exact Resource URL shown on the DevTools Resource Details page. The owner in that URL may be another user when a resource is shared with you.

Disable cloud forwarding

alshival.configure(cloud_level='NONE')

Accepted cloud_level values: DEBUG, INFO, WARNING, ERROR, ALERT, NONE.

Logging

import logging
import alshival

logging.basicConfig(level=logging.INFO)

alshival.log.info('local logging flow')
alshival.log.warning('forwarded when cloud_level <= WARNING')
alshival.log.error('forwarded when cloud_level <= ERROR')
alshival.log.alert('high-priority incident')

MCP tool descriptors

import alshival
from openai import OpenAI

client = OpenAI()

tools = [
    alshival.mcp,
    alshival.mcp.github,
]

response = client.responses.create(
    model='gpt-5.2',
    input='Review deployment risks in my repos.',
    tools=tools,
)

Notes

ALSHIVAL_RESOURCE is the preferred targeting input.
ALSHIVAL_DEBUG=true enables SDK diagnostics to stderr.
Cloud-level filtering only controls cloud forwarding. It does not suppress your local logger output.