Agent Zero AI framework
Document and verify greeting history and log ownership
Alessandrobf31f47c7bb0
diff --git a/extensions/python/agent_init/AGENTS.md b/extensions/python/agent_init/AGENTS.mdindex 7d0adbcae..e958fa34c 100644--- a/extensions/python/agent_init/AGENTS.md+++ b/extensions/python/agent_init/AGENTS.md@@ -15,7 +15,7 @@- Keep a placeholder user turn (`fw.initial_user_message.md`) ahead of the AI greeting (`fw.initial_message.md`) so `output_langchain` never pops the greeting as a leading `AIMessage`; do not remove either prompt without replacing the turn-order guarantee.- Minify the initial AI message JSON before storage; preserve raw text as the displayed greeting when JSON parsing fails.- `_10_initial_message.py` passes `LLMResult.non_llm()` to `hist_add_ai_response` so the Responses-API state seam runs uniformly; the sentinel carries no `response_id` and marks `mode=""`/`state="off"` so stored metadata does not claim a Responses-API turn.-- The synthetic `Hello!` user turn and the AI greeting are intentionally hidden: they have no matching UI log item (branching links log↔history by ID, but these seeding turns predate the log). They serve as a few-shot example of the expected user→AI turn ordering so `output_langchain` does not pop the greeting as a leading `AIMessage`. Do not add log items for seeding turns without replacing the few-shot guarantee.+- The synthetic `Hello!` user turn has no UI log. The assistant greeting is displayed as a response log with the same ID as its history message, preserving the log/history link for branching. Keep the user-first seed without duplicating the visible greeting on reload.## Work Guidancediff --git a/tests/test_initial_message.py b/tests/test_initial_message.pynew file mode 100644index 000000000..25fff908b--- /dev/null+++ b/tests/test_initial_message.py@@ -0,0 +1,52 @@+import json+from types import SimpleNamespace++import pytest+from langchain_core.messages import HumanMessage, AIMessage++from extensions.python.agent_init._10_initial_message import InitialMessage+from helpers import history+from helpers.llm_result import result_from_metadata+from helpers.log import Log++[email protected]("greeting", [+ json.dumps({"tool_name": "response", "tool_args": {"text": "Welcome."}}, indent=2),+ "Welcome to this profile.",+])+def test_initial_greeting_keeps_user_first_order_and_linked_log_after_reload(greeting):+ messages = []+ log = Log()++ def add_ai(content, llm_result):+ message = history.Message(True, content, metadata=llm_result.metadata())+ messages.append(message)+ return message++ agent = SimpleNamespace(+ number=0,+ context=SimpleNamespace(log=log),+ read_prompt=lambda name: "Hello!" if name == "fw.initial_user_message.md" else greeting,+ hist_add_user_message=lambda message: messages.append(history.Message(False, message.message)),+ hist_add_ai_response=add_ai,+ )+ extension = InitialMessage(agent)+ extension.execute()+ extension.execute()+ assert len(messages) == 2+ assert len(log.logs) == 1+ assert log.logs[0].id == messages[1].id+ assert log.logs[0].id != messages[0].id++ reloaded = [history.Message.from_dict(message.to_dict(), history=None) for message in messages]+ rendered = history.output_langchain([item for message in reloaded for item in message.output()])+ assert [type(item) for item in rendered] == [HumanMessage, AIMessage]+ assert rendered[0].content == "Hello!"+ assert rendered[1].content == messages[1].content+ result = result_from_metadata(reloaded[1].metadata)+ assert result.mode == "" and result.state == "off" and not result.response_id+ if greeting.startswith("{"):+ assert "\n" not in rendered[1].content+ assert log.logs[0].content == "Welcome."+ else:+ assert log.logs[0].content == greeting