Richard Tang
786 lines
26 KiB
Python
786 lines
26 KiB
Python
"""Tests for the queen global memory system (reflection + recall)."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import asyncio
|
|
import json
|
|
import time
|
|
from pathlib import Path
|
|
from types import SimpleNamespace
|
|
from unittest.mock import AsyncMock, MagicMock
|
|
|
|
import pytest
|
|
|
|
from framework.agents.queen import queen_memory_v2 as qm
|
|
from framework.agents.queen.recall_selector import (
|
|
build_scoped_recall_blocks,
|
|
format_recall_injection,
|
|
select_memories,
|
|
)
|
|
from framework.orchestrator.prompting import build_system_prompt_for_node_context
|
|
from framework.server.queen_orchestrator import initialize_memory_scopes
|
|
from framework.tools.queen_lifecycle_tools import QueenPhaseState
|
|
|
|
|
|
def _make_litellm_response(tool_calls: list[dict] | None = None, content: str = ""):
|
|
"""Build a mock that mirrors litellm ModelResponse structure."""
|
|
if tool_calls:
|
|
tc_objects = []
|
|
for tc in tool_calls:
|
|
fn = SimpleNamespace(
|
|
name=tc["name"],
|
|
arguments=json.dumps(tc.get("input", {})),
|
|
)
|
|
tc_objects.append(SimpleNamespace(id=tc["id"], function=fn))
|
|
message = SimpleNamespace(tool_calls=tc_objects)
|
|
else:
|
|
message = SimpleNamespace(tool_calls=None)
|
|
raw = SimpleNamespace(choices=[SimpleNamespace(message=message)])
|
|
return MagicMock(content=content, raw_response=raw)
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# parse_frontmatter
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_parse_frontmatter_valid():
|
|
text = "---\nname: foo\ntype: profile\ndescription: bar baz\n---\ncontent"
|
|
fm = qm.parse_frontmatter(text)
|
|
assert fm == {"name": "foo", "type": "profile", "description": "bar baz"}
|
|
|
|
|
|
def test_parse_frontmatter_missing():
|
|
assert qm.parse_frontmatter("no frontmatter here") == {}
|
|
|
|
|
|
def test_parse_frontmatter_empty():
|
|
assert qm.parse_frontmatter("") == {}
|
|
|
|
|
|
def test_parse_frontmatter_broken_yaml():
|
|
text = "---\n: bad\nno colon\n---\n"
|
|
fm = qm.parse_frontmatter(text)
|
|
assert fm == {}
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# parse_global_memory_category
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_parse_global_memory_category_valid():
|
|
assert qm.parse_global_memory_category("profile") == "profile"
|
|
assert qm.parse_global_memory_category("preference") == "preference"
|
|
assert qm.parse_global_memory_category("environment") == "environment"
|
|
assert qm.parse_global_memory_category("feedback") == "feedback"
|
|
|
|
|
|
def test_parse_global_memory_category_case_insensitive():
|
|
assert qm.parse_global_memory_category("Profile") == "profile"
|
|
assert qm.parse_global_memory_category(" FEEDBACK ") == "feedback"
|
|
|
|
|
|
def test_parse_global_memory_category_invalid():
|
|
assert qm.parse_global_memory_category("goal") is None
|
|
assert qm.parse_global_memory_category("unknown") is None
|
|
assert qm.parse_global_memory_category(None) is None
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# MemoryFile.from_path
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_memory_file_from_path(tmp_path: Path):
|
|
f = tmp_path / "test.md"
|
|
f.write_text("---\nname: test\ntype: profile\ndescription: a test\n---\nbody\n")
|
|
mf = qm.MemoryFile.from_path(f)
|
|
assert mf.filename == "test.md"
|
|
assert mf.name == "test"
|
|
assert mf.type == "profile"
|
|
assert mf.description == "a test"
|
|
assert mf.mtime > 0
|
|
|
|
|
|
def test_memory_file_from_path_no_frontmatter(tmp_path: Path):
|
|
f = tmp_path / "bare.md"
|
|
f.write_text("just plain text\n")
|
|
mf = qm.MemoryFile.from_path(f)
|
|
assert mf.name is None
|
|
assert mf.type is None
|
|
assert mf.description is None
|
|
assert "just plain text" in mf.header_lines
|
|
|
|
|
|
def test_memory_file_from_path_missing(tmp_path: Path):
|
|
f = tmp_path / "missing.md"
|
|
mf = qm.MemoryFile.from_path(f)
|
|
assert mf.filename == "missing.md"
|
|
assert mf.name is None
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# scan_memory_files
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_scan_memory_files(tmp_path: Path):
|
|
(tmp_path / "a.md").write_text("---\nname: a\n---\n")
|
|
time.sleep(0.01)
|
|
(tmp_path / "b.md").write_text("---\nname: b\n---\n")
|
|
(tmp_path / ".hidden.md").write_text("---\nname: hidden\n---\n")
|
|
(tmp_path / "not-md.txt").write_text("ignored")
|
|
|
|
files = qm.scan_memory_files(tmp_path)
|
|
names = [f.filename for f in files]
|
|
assert "a.md" in names
|
|
assert "b.md" in names
|
|
assert ".hidden.md" not in names
|
|
assert "not-md.txt" not in names
|
|
# Newest first.
|
|
assert names[0] == "b.md"
|
|
|
|
|
|
def test_scan_memory_files_cap(tmp_path: Path):
|
|
for i in range(210):
|
|
(tmp_path / f"mem-{i:04d}.md").write_text(f"---\nname: m{i}\n---\n")
|
|
files = qm.scan_memory_files(tmp_path)
|
|
assert len(files) == qm.MAX_FILES
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# format_memory_manifest
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_format_memory_manifest():
|
|
files = [
|
|
qm.MemoryFile(
|
|
filename="a.md",
|
|
path=Path("a.md"),
|
|
name="a",
|
|
type="profile",
|
|
description="desc a",
|
|
mtime=time.time(),
|
|
),
|
|
qm.MemoryFile(
|
|
filename="b.md",
|
|
path=Path("b.md"),
|
|
name="b",
|
|
type=None,
|
|
description=None,
|
|
mtime=0.0,
|
|
),
|
|
]
|
|
manifest = qm.format_memory_manifest(files)
|
|
assert "[profile] a.md" in manifest
|
|
assert "desc a" in manifest
|
|
assert "[unknown] b.md" in manifest
|
|
assert "(no description)" in manifest
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# init_memory_dir
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_init_memory_dir(tmp_path: Path):
|
|
mem_dir = tmp_path / "memories"
|
|
qm.init_memory_dir(mem_dir)
|
|
assert mem_dir.is_dir()
|
|
|
|
|
|
def test_initialize_memory_scopes_uses_queen_memory_dir(tmp_path: Path, monkeypatch):
|
|
global_dir = tmp_path / "memories" / "global"
|
|
queen_dir = tmp_path / "memories" / "agents" / "queens" / "queen_technology"
|
|
|
|
monkeypatch.setattr(qm, "global_memory_dir", lambda: global_dir)
|
|
monkeypatch.setattr(qm, "queen_memory_dir", lambda queen_name="default": queen_dir)
|
|
|
|
session = SimpleNamespace(queen_name="queen_technology")
|
|
phase = QueenPhaseState()
|
|
|
|
resolved_global, resolved_queen = initialize_memory_scopes(session, phase)
|
|
|
|
assert resolved_global == global_dir
|
|
assert resolved_queen == queen_dir
|
|
assert phase.global_memory_dir == global_dir
|
|
assert phase.queen_memory_dir == queen_dir
|
|
assert global_dir.is_dir()
|
|
assert queen_dir.is_dir()
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# recall_selector
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_select_memories_empty_dir(tmp_path: Path):
|
|
llm = AsyncMock()
|
|
result = await select_memories("hello", llm, memory_dir=tmp_path)
|
|
assert result == []
|
|
llm.acomplete.assert_not_called()
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_select_memories_with_files(tmp_path: Path):
|
|
(tmp_path / "a.md").write_text("---\nname: a\ndescription: about A\ntype: profile\n---\nbody")
|
|
(tmp_path / "b.md").write_text(
|
|
"---\nname: b\ndescription: about B\ntype: preference\n---\nbody"
|
|
)
|
|
|
|
llm = AsyncMock()
|
|
llm.acomplete.return_value = MagicMock(content=json.dumps({"selected_memories": ["a.md"]}))
|
|
|
|
result = await select_memories("tell me about A", llm, memory_dir=tmp_path)
|
|
assert result == ["a.md"]
|
|
llm.acomplete.assert_called_once()
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_select_memories_error_returns_empty(tmp_path: Path):
|
|
(tmp_path / "a.md").write_text("---\nname: a\n---\nbody")
|
|
|
|
llm = AsyncMock()
|
|
llm.acomplete.side_effect = RuntimeError("LLM down")
|
|
|
|
result = await select_memories("hello", llm, memory_dir=tmp_path)
|
|
assert result == []
|
|
|
|
|
|
def test_format_recall_injection(tmp_path: Path):
|
|
(tmp_path / "a.md").write_text("---\nname: a\n---\nbody of a")
|
|
result = format_recall_injection(["a.md"], memory_dir=tmp_path)
|
|
assert "Global Memories" in result
|
|
assert "body of a" in result
|
|
|
|
|
|
def test_format_recall_injection_custom_label(tmp_path: Path):
|
|
(tmp_path / "a.md").write_text("---\nname: a\n---\nbody of a")
|
|
result = format_recall_injection(["a.md"], memory_dir=tmp_path, label="Queen Memories: queen_technology")
|
|
assert "Queen Memories: queen_technology" in result
|
|
assert "body of a" in result
|
|
|
|
|
|
def test_format_recall_injection_empty():
|
|
assert format_recall_injection([]) == ""
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_build_scoped_recall_blocks_includes_global_and_queen(tmp_path: Path):
|
|
global_dir = tmp_path / "global"
|
|
queen_dir = tmp_path / "queen"
|
|
global_dir.mkdir()
|
|
queen_dir.mkdir()
|
|
(global_dir / "shared.md").write_text("---\nname: shared\n---\nshared body")
|
|
(queen_dir / "shared.md").write_text("---\nname: shared\n---\nqueen body")
|
|
|
|
llm = AsyncMock()
|
|
llm.acomplete.side_effect = [
|
|
MagicMock(content=json.dumps({"selected_memories": ["shared.md"]})),
|
|
MagicMock(content=json.dumps({"selected_memories": ["shared.md"]})),
|
|
]
|
|
|
|
global_block, queen_block = await build_scoped_recall_blocks(
|
|
"help me",
|
|
llm,
|
|
global_memory_dir=global_dir,
|
|
queen_memory_dir=queen_dir,
|
|
queen_id="queen_technology",
|
|
)
|
|
|
|
assert "Global Memories" in global_block
|
|
assert "shared body" in global_block
|
|
assert "Queen Memories: queen_technology" in queen_block
|
|
assert "queen body" in queen_block
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_build_scoped_recall_blocks_tolerates_empty_scope(tmp_path: Path):
|
|
global_dir = tmp_path / "global"
|
|
queen_dir = tmp_path / "queen"
|
|
global_dir.mkdir()
|
|
queen_dir.mkdir()
|
|
(global_dir / "a.md").write_text("---\nname: a\n---\nglobal body")
|
|
|
|
llm = AsyncMock()
|
|
llm.acomplete.return_value = MagicMock(content=json.dumps({"selected_memories": ["a.md"]}))
|
|
|
|
global_block, queen_block = await build_scoped_recall_blocks(
|
|
"help me",
|
|
llm,
|
|
global_memory_dir=global_dir,
|
|
queen_memory_dir=queen_dir,
|
|
queen_id="queen_technology",
|
|
)
|
|
|
|
assert "Global Memories" in global_block
|
|
assert queen_block == ""
|
|
llm.acomplete.assert_called_once()
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# reflection_agent
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_short_reflection(tmp_path: Path):
|
|
"""Short reflection reads messages and writes a global memory file via LLM tools."""
|
|
from framework.agents.queen.reflection_agent import run_short_reflection
|
|
|
|
parts_dir = tmp_path / "session" / "conversations" / "parts"
|
|
parts_dir.mkdir(parents=True)
|
|
for i in range(3):
|
|
role = "user" if i % 2 == 0 else "assistant"
|
|
(parts_dir / f"{i:010d}.json").write_text(
|
|
json.dumps({"role": role, "content": f"message {i}"})
|
|
)
|
|
|
|
mem_dir = tmp_path / "global_memory"
|
|
mem_dir.mkdir()
|
|
|
|
llm = AsyncMock()
|
|
llm.acomplete.side_effect = [
|
|
# Turn 1: LLM writes a global memory file
|
|
_make_litellm_response(
|
|
tool_calls=[
|
|
{
|
|
"id": "tc_1",
|
|
"name": "write_memory_file",
|
|
"input": {
|
|
"filename": "user-likes-tests.md",
|
|
"content": (
|
|
"---\nname: user-likes-tests\n"
|
|
"type: preference\n"
|
|
"description: User values thorough testing\n"
|
|
"---\nObserved emphasis on test coverage."
|
|
),
|
|
},
|
|
}
|
|
]
|
|
),
|
|
# Turn 2: done
|
|
_make_litellm_response(content="Done reflecting."),
|
|
]
|
|
|
|
session_dir = tmp_path / "session"
|
|
await run_short_reflection(session_dir, llm, memory_dir=mem_dir)
|
|
|
|
written = mem_dir / "user-likes-tests.md"
|
|
assert written.exists()
|
|
assert "user-likes-tests" in written.read_text()
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_queen_short_reflection_writes_only_queen_scope(tmp_path: Path):
|
|
"""Queen short reflection writes to queen memory without touching global memory."""
|
|
from framework.agents.queen.reflection_agent import run_queen_short_reflection
|
|
|
|
parts_dir = tmp_path / "session" / "conversations" / "parts"
|
|
parts_dir.mkdir(parents=True)
|
|
for i in range(3):
|
|
role = "user" if i % 2 == 0 else "assistant"
|
|
(parts_dir / f"{i:010d}.json").write_text(
|
|
json.dumps({"role": role, "content": f"message {i}"})
|
|
)
|
|
|
|
global_dir = tmp_path / "global_memory"
|
|
queen_dir = tmp_path / "queen_memory"
|
|
global_dir.mkdir()
|
|
queen_dir.mkdir()
|
|
|
|
llm = AsyncMock()
|
|
llm.acomplete.side_effect = [
|
|
_make_litellm_response(
|
|
tool_calls=[
|
|
{
|
|
"id": "tc_1",
|
|
"name": "write_memory_file",
|
|
"input": {
|
|
"filename": "technology-workflow.md",
|
|
"content": (
|
|
"---\nname: technology-workflow\n"
|
|
"type: preference\n"
|
|
"description: User prefers implementation-first technical help\n"
|
|
"---\nFor technical work, the user wants concrete implementation details."
|
|
),
|
|
},
|
|
}
|
|
]
|
|
),
|
|
_make_litellm_response(content="Done reflecting."),
|
|
]
|
|
|
|
await run_queen_short_reflection(
|
|
tmp_path / "session",
|
|
llm,
|
|
"queen_technology",
|
|
queen_dir,
|
|
)
|
|
|
|
assert (queen_dir / "technology-workflow.md").exists()
|
|
assert list(global_dir.glob("*.md")) == []
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_unified_short_reflection_can_write_both_scopes_in_one_loop(tmp_path: Path):
|
|
"""Unified short reflection can place memories in both scopes in one pass."""
|
|
from framework.agents.queen.reflection_agent import run_unified_short_reflection
|
|
|
|
parts_dir = tmp_path / "session" / "conversations" / "parts"
|
|
parts_dir.mkdir(parents=True)
|
|
for i in range(3):
|
|
role = "user" if i % 2 == 0 else "assistant"
|
|
(parts_dir / f"{i:010d}.json").write_text(
|
|
json.dumps({"role": role, "content": f"message {i}"})
|
|
)
|
|
|
|
global_dir = tmp_path / "global_memory"
|
|
queen_dir = tmp_path / "queen_memory"
|
|
global_dir.mkdir()
|
|
queen_dir.mkdir()
|
|
|
|
llm = AsyncMock()
|
|
llm.acomplete.side_effect = [
|
|
_make_litellm_response(
|
|
tool_calls=[
|
|
{
|
|
"id": "tc_1",
|
|
"name": "write_memory_file",
|
|
"input": {
|
|
"scope": "global",
|
|
"filename": "user-profile.md",
|
|
"content": (
|
|
"---\nname: User Profile\n"
|
|
"type: profile\n"
|
|
"description: Shared user profile\n"
|
|
"---\nShared profile body."
|
|
),
|
|
},
|
|
},
|
|
{
|
|
"id": "tc_2",
|
|
"name": "write_memory_file",
|
|
"input": {
|
|
"scope": "queen",
|
|
"filename": "technology-preferences.md",
|
|
"content": (
|
|
"---\nname: technology-preferences\n"
|
|
"type: preference\n"
|
|
"description: Technical execution preferences\n"
|
|
"---\nThe user wants implementation-first technical answers."
|
|
),
|
|
},
|
|
},
|
|
]
|
|
),
|
|
_make_litellm_response(content="Done reflecting."),
|
|
]
|
|
|
|
await run_unified_short_reflection(
|
|
tmp_path / "session",
|
|
llm,
|
|
global_memory_dir=global_dir,
|
|
queen_memory_dir=queen_dir,
|
|
queen_id="queen_technology",
|
|
)
|
|
|
|
assert (global_dir / "user-profile.md").exists()
|
|
assert (queen_dir / "technology-preferences.md").exists()
|
|
assert llm.acomplete.await_count == 2
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_short_reflection_rejects_non_global_types(tmp_path: Path):
|
|
"""Reflection agent rejects memory types not in GLOBAL_MEMORY_CATEGORIES."""
|
|
from framework.agents.queen.reflection_agent import _execute_tool
|
|
|
|
mem_dir = tmp_path / "global_memory"
|
|
mem_dir.mkdir()
|
|
|
|
result = _execute_tool(
|
|
"write_memory_file",
|
|
{
|
|
"filename": "bad-type.md",
|
|
"content": "---\nname: bad\ntype: goal\n---\nbody",
|
|
},
|
|
mem_dir,
|
|
)
|
|
assert "ERROR" in result
|
|
assert not (mem_dir / "bad-type.md").exists()
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_long_reflection(tmp_path: Path):
|
|
"""Long reflection reads all memories and can merge/delete them."""
|
|
from framework.agents.queen.reflection_agent import run_long_reflection
|
|
|
|
mem_dir = tmp_path / "global_memory"
|
|
mem_dir.mkdir()
|
|
(mem_dir / "dup-a.md").write_text(
|
|
"---\nname: dup-a\ntype: profile\ndescription: profile A\n---\nProfile A details."
|
|
)
|
|
(mem_dir / "dup-b.md").write_text(
|
|
"---\nname: dup-b\ntype: profile\ndescription: profile A dup\n---\nSame profile A."
|
|
)
|
|
|
|
llm = AsyncMock()
|
|
llm.acomplete.side_effect = [
|
|
_make_litellm_response(
|
|
tool_calls=[
|
|
{"id": "tc_1", "name": "list_memory_files", "input": {}},
|
|
]
|
|
),
|
|
_make_litellm_response(
|
|
tool_calls=[
|
|
{
|
|
"id": "tc_2",
|
|
"name": "write_memory_file",
|
|
"input": {
|
|
"filename": "dup-a.md",
|
|
"content": (
|
|
"---\nname: dup-a\ntype: profile\n"
|
|
"description: profile A (merged)\n"
|
|
"---\nProfile A details. Also same profile A."
|
|
),
|
|
},
|
|
},
|
|
{
|
|
"id": "tc_3",
|
|
"name": "delete_memory_file",
|
|
"input": {"filename": "dup-b.md"},
|
|
},
|
|
]
|
|
),
|
|
_make_litellm_response(content="Housekeeping complete."),
|
|
]
|
|
|
|
await run_long_reflection(llm, memory_dir=mem_dir)
|
|
|
|
assert not (mem_dir / "dup-b.md").exists()
|
|
assert (mem_dir / "dup-a.md").exists()
|
|
assert "merged" in (mem_dir / "dup-a.md").read_text()
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_subscribe_reflection_triggers_runs_housekeeping_for_both_scopes(
|
|
tmp_path: Path,
|
|
monkeypatch,
|
|
):
|
|
from framework.agents.queen import reflection_agent as ra
|
|
from framework.host.event_bus import AgentEvent, EventBus, EventType
|
|
|
|
bus = EventBus()
|
|
session_dir = tmp_path / "session"
|
|
global_dir = tmp_path / "global"
|
|
queen_dir = tmp_path / "queen"
|
|
global_dir.mkdir()
|
|
queen_dir.mkdir()
|
|
llm = AsyncMock()
|
|
|
|
unified_short = AsyncMock()
|
|
unified_long = AsyncMock()
|
|
|
|
monkeypatch.setattr(ra, "run_unified_short_reflection", unified_short)
|
|
monkeypatch.setattr(ra, "run_unified_long_reflection", unified_long)
|
|
|
|
sub_ids = await ra.subscribe_reflection_triggers(
|
|
bus,
|
|
session_dir,
|
|
llm,
|
|
global_memory_dir=global_dir,
|
|
queen_memory_dir=queen_dir,
|
|
queen_id="queen_technology",
|
|
)
|
|
|
|
for _ in range(5):
|
|
await bus.publish(
|
|
AgentEvent(
|
|
type=EventType.LLM_TURN_COMPLETE,
|
|
stream_id="queen",
|
|
data={"stop_reason": "stop"},
|
|
)
|
|
)
|
|
|
|
await asyncio.sleep(0.05)
|
|
|
|
assert len(sub_ids) == 2
|
|
assert unified_short.await_count == 5
|
|
unified_long.assert_awaited_once_with(
|
|
llm,
|
|
global_memory_dir=global_dir,
|
|
queen_memory_dir=queen_dir,
|
|
queen_id="queen_technology",
|
|
)
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_shutdown_reflection_writes_global_and_queen_scope(tmp_path: Path):
|
|
from framework.agents.queen.reflection_agent import run_shutdown_reflection
|
|
|
|
parts_dir = tmp_path / "session" / "conversations" / "parts"
|
|
parts_dir.mkdir(parents=True)
|
|
for i in range(3):
|
|
role = "user" if i % 2 == 0 else "assistant"
|
|
(parts_dir / f"{i:010d}.json").write_text(
|
|
json.dumps({"role": role, "content": f"message {i}"})
|
|
)
|
|
|
|
global_dir = tmp_path / "global_memory"
|
|
queen_dir = tmp_path / "queen_memory"
|
|
global_dir.mkdir()
|
|
queen_dir.mkdir()
|
|
|
|
llm = AsyncMock()
|
|
llm.acomplete.side_effect = [
|
|
_make_litellm_response(
|
|
tool_calls=[
|
|
{
|
|
"id": "tc_1",
|
|
"name": "write_memory_file",
|
|
"input": {
|
|
"scope": "global",
|
|
"filename": "user-profile.md",
|
|
"content": (
|
|
"---\nname: User Profile\n"
|
|
"type: profile\n"
|
|
"description: Shared user profile\n"
|
|
"---\nShared profile body."
|
|
),
|
|
},
|
|
},
|
|
{
|
|
"id": "tc_2",
|
|
"name": "write_memory_file",
|
|
"input": {
|
|
"scope": "queen",
|
|
"filename": "technology-preferences.md",
|
|
"content": (
|
|
"---\nname: technology-preferences\n"
|
|
"type: preference\n"
|
|
"description: Technical execution preferences\n"
|
|
"---\nThe user wants implementation-first technical answers."
|
|
),
|
|
},
|
|
},
|
|
]
|
|
),
|
|
_make_litellm_response(content="Done reflecting."),
|
|
]
|
|
|
|
await run_shutdown_reflection(
|
|
tmp_path / "session",
|
|
llm,
|
|
global_memory_dir_override=global_dir,
|
|
queen_memory_dir=queen_dir,
|
|
queen_id="queen_technology",
|
|
)
|
|
|
|
assert (global_dir / "user-profile.md").exists()
|
|
assert (queen_dir / "technology-preferences.md").exists()
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Path traversal prevention
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_path_traversal_read(tmp_path: Path):
|
|
from framework.agents.queen.reflection_agent import _execute_tool
|
|
|
|
(tmp_path / "safe.md").write_text("safe content")
|
|
result = _execute_tool("read_memory_file", {"filename": "../../etc/passwd"}, tmp_path)
|
|
assert "ERROR" in result
|
|
|
|
|
|
def test_path_traversal_write(tmp_path: Path):
|
|
from framework.agents.queen.reflection_agent import _execute_tool
|
|
|
|
result = _execute_tool(
|
|
"write_memory_file",
|
|
{"filename": "../escape.md", "content": "---\nname: evil\n---\nbad"},
|
|
tmp_path,
|
|
)
|
|
assert "ERROR" in result
|
|
assert not (tmp_path.parent / "escape.md").exists()
|
|
|
|
|
|
def test_safe_path_accepted(tmp_path: Path):
|
|
from framework.agents.queen.reflection_agent import _execute_tool
|
|
|
|
result = _execute_tool(
|
|
"write_memory_file",
|
|
{"filename": "good-file.md", "content": "---\nname: good\ntype: profile\n---\ncontent"},
|
|
tmp_path,
|
|
)
|
|
assert "Wrote" in result
|
|
assert (tmp_path / "good-file.md").exists()
|
|
|
|
result = _execute_tool("read_memory_file", {"filename": "good-file.md"}, tmp_path)
|
|
assert "content" in result
|
|
|
|
result = _execute_tool("delete_memory_file", {"filename": "good-file.md"}, tmp_path)
|
|
assert "Deleted" in result
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# system prompt integration
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_build_system_prompt_injects_dynamic_memory():
|
|
ctx = SimpleNamespace(
|
|
identity_prompt="Identity",
|
|
node_spec=SimpleNamespace(
|
|
system_prompt="Focus", node_type="event_loop", output_keys=["out"]
|
|
),
|
|
narrative="Narrative",
|
|
accounts_prompt="",
|
|
skills_catalog_prompt="",
|
|
protocols_prompt="",
|
|
memory_prompt="",
|
|
dynamic_memory_provider=lambda: "--- Global Memories ---\nremember this",
|
|
is_subagent_mode=False,
|
|
)
|
|
|
|
prompt = build_system_prompt_for_node_context(ctx)
|
|
assert "Global Memories" in prompt
|
|
assert "remember this" in prompt
|
|
|
|
|
|
def test_queen_phase_state_appends_global_memory_block():
|
|
phase = QueenPhaseState(
|
|
prompt_building="base prompt",
|
|
_cached_global_recall_block="--- Global Memories ---\nglobal stuff",
|
|
)
|
|
|
|
prompt = phase.get_current_prompt()
|
|
assert "base prompt" in prompt
|
|
assert "Global Memories" in prompt
|
|
assert "global stuff" in prompt
|
|
|
|
|
|
def test_queen_phase_state_appends_queen_memory_block():
|
|
phase = QueenPhaseState(
|
|
prompt_building="base prompt",
|
|
_cached_global_recall_block="--- Global Memories ---\nglobal stuff",
|
|
_cached_queen_recall_block="--- Queen Memories: queen_technology ---\nqueen stuff",
|
|
)
|
|
|
|
prompt = phase.get_current_prompt()
|
|
assert "base prompt" in prompt
|
|
assert "Global Memories" in prompt
|
|
assert "Queen Memories: queen_technology" in prompt
|
|
assert "queen stuff" in prompt
|
|
|
|
|
|
def test_queen_phase_state_prompt_without_memory():
|
|
phase = QueenPhaseState(prompt_building="base prompt")
|
|
|
|
prompt = phase.get_current_prompt()
|
|
assert "base prompt" in prompt
|
|
assert "Global Memories" not in prompt
|