Open-source AI coworkers that each get a computer of their own: a browser, files and tools, with every action decided before it happens and recorded after. Bring any AG-UI agent.
"""Exercise learned skill reads through the real bundled framework."""
import sys
from pathlib import Path
ROOT = Path(__file__).resolve().parents[2]
sys.path.insert(0, str(ROOT / "tests"))
sys.path.insert(0, str(ROOT / "agent-adk"))
from learning_harness_contract import assert_learning_delivery
def test_learned_catalog_and_both_read_tools_reach_the_model(monkeypatch):
assert_learning_delivery(monkeypatch, "agent-adk", "/")
def test_partial_tool_arguments_wait_without_dropping_neighboring_text(monkeypatch):
import asyncio
import importlib
from google.adk.models.lite_llm import LiteLlm
from google.adk.models.llm_response import LlmResponse
from google.genai import types
monkeypatch.setenv("BOT_MODEL", "gpt-4.1-mini")
main = importlib.import_module("src.main")
async def provider(self, request, stream=False):
yield LlmResponse(partial=True, content=types.Content(role="model", parts=[
types.Part(text="Looking up the policy."),
types.Part(function_call=types.FunctionCall(id="call", name="copilotkit_load_skill", will_continue=True)),
]))
yield LlmResponse(partial=False, content=types.Content(role="model", parts=[
types.Part(function_call=types.FunctionCall(id="call", name="copilotkit_load_skill", args={"skill_name": "refunds"})),
]))
monkeypatch.setattr(LiteLlm, "generate_content_async", provider)
async def collect():
model = main.CompleteToolCallsLiteLlm(model="openai/gpt-4.1-mini")
return [response async for response in model.generate_content_async(None, True)]
responses = asyncio.run(collect())
assert responses[0].content.parts == [types.Part(text="Looking up the policy.")]
assert responses[1].content.parts[0].function_call.args == {"skill_name": "refunds"}
assert responses[1].content.parts[0].function_call.id == "call"