Put work items in the state store, and bind the log to them (M1.1)

Development Plan M1.1, as discussed.

- Schema v4: a work_items table (id, scope, title, status, outcome,
  position, data) and an agent row whose active_item points at one
  (NULL: idle). Canonical events and intents carry their work item, and
  events their actor. Older stores are refused with a clear message:
  there was no production data, so no migration (the mechanism stays).
- The log is reached through a Transcript bound to one work item and
  one actor: store.transcript(item, actor) reads only that item and
  writes with both. The planner, executor, delegation and expert desk
  get one per call from the orchestrator instead of holding the store,
  so a component can neither read another task's events (the planner
  read every earlier task's, #78) nor write without attribution.
  store.all_events() is kept for readers that span items (evaluation
  capture, dumps). Intents are begun through the transcript or with
  work_item=, so each belongs to its item.
- The orchestrator opens a work item on 'use', closes it with its
  outcome (completed or abandoned), and carries what the operator said
  while idle into the new item as copies with their origin, so a task
  sent before 'use' is still planned from (D1). control.py keeps its
  state machine through get_state()/transition(), now over the active
  work item, until M2.6 retires it.
- nisse_dev.dump shows work items and each event's item and actor.

Tests: test_work_items.py (scoping, attribution, intent ownership,
restart, outcomes, carried messages, a second task never seeing the
first); the suites updated for the per-call transcript.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PPQ14L3LikyefCMFBqEbKE

tests/integration/test_crash_durability.py

@@ -21,7 +21,7 @@ from nisse_dev.state import StateStore
 store = StateStore(sys.argv[1])
 store.transition("EXECUTING", cause="test", project="demo", data={"milestone": 2})
 store.begin_intent("push_branch", {"branch": "milestone-2"}, cause="test")
-store.append_event(UserMessage(text="start milestone 2", source="m:1"))
+store.transcript("t", "operator").append(UserMessage(text="start milestone 2", source="m:1"))
 os._exit(1)  # die without closing anything, as a power cut would
 """
 
@@ -48,5 +48,5 @@ def test_state_survives_process_death(tmp_path: Path) -> None:
         [pending] = store.pending_intents()
         assert pending.payload == {"branch": "milestone-2"}
 
-        [stored] = store.read_events()
+        [stored] = store.all_events()
         assert stored.event.text == "start milestone 2"

tests/integration/test_delegation_live.py

@@ -45,11 +45,13 @@ def build(
 def test_delegate_completes_a_scoped_subtask(
     tmp_path: Path, passthrough: AnthropicPassthroughClient, live_model: str
 ) -> None:
-    delegation, repo, _store = build(
+    delegation, repo, store = build(
         tmp_path, Caps(turns=12, wall_clock_s=300.0), passthrough, live_model
     )
 
-    result = delegation.run(brief=BRIEF, branch="live/delegation-1")
+    result = delegation.run(
+        store.transcript("t", "helper"), brief=BRIEF, branch="live/delegation-1"
+    )
 
     assert result.usable, f"delegation did not hand back: {result.status} {result.note}"
     assert (repo.root / "slugify.py").exists()
@@ -75,11 +77,13 @@ def test_a_cap_breach_discards_the_work(
 ) -> None:
     # One turn cannot finish the brief, so this exercises the abort path
     # against a real model rather than a scripted one.
-    delegation, repo, _ = build(
+    delegation, repo, store = build(
         tmp_path, Caps(turns=1, wall_clock_s=120.0), passthrough, live_model
     )
 
-    result = delegation.run(brief=BRIEF, branch="live/delegation-2")
+    result = delegation.run(
+        store.transcript("t", "helper"), brief=BRIEF, branch="live/delegation-2"
+    )
 
     assert not result.usable
     assert result.status == "capped"

tests/integration/test_expert_live.py

@@ -26,7 +26,7 @@ def desk(tmp_path: Path, litellm: LiteLLMClient, live_model: str):
         store = StateStore(
             tmp_path / "state.db", oplog=OpLog("nisse-dev", io.StringIO())
         )
-        return ExpertDesk(litellm, store, roles, budget=budget), store
+        return ExpertDesk(litellm, roles, budget=budget), store
 
     return make
 
@@ -35,8 +35,9 @@ def test_live_consult_parses_against_the_local_schema(desk) -> None:
     expert, store = desk()
 
     answer = expert.ask(
+        store.transcript("t", "expert"),
         "For a single-process agent storing a few thousand rows of task state, "
-        "is SQLite sufficient, or is PostgreSQL warranted? Answer briefly."
+        "is SQLite sufficient, or is PostgreSQL warranted? Answer briefly.",
     )
 
     assert answer.confidence in {"high", "medium", "low"}
@@ -48,7 +49,10 @@ def test_live_consult_parses_against_the_local_schema(desk) -> None:
 
 
 def test_live_budget_stops_a_circling_caller(desk) -> None:
-    expert, _ = desk(budget=1)
-    expert.ask("Name one advantage of WAL mode in SQLite. Be brief.")
+    expert, store = desk(budget=1)
+    expert.ask(
+        store.transcript("t", "expert"),
+        "Name one advantage of WAL mode in SQLite. Be brief.",
+    )
     with pytest.raises(BudgetExhausted):
-        expert.ask("And another?")
+        expert.ask(store.transcript("t", "expert"), "And another?")

tests/integration/test_milestone_live.py

@@ -72,10 +72,10 @@ def test_plan_approve_execute_produces_a_real_file(
     orchestrator = Orchestrator(
         store,
         transport,
-        LLMPlanningAgent(client, store, roles),
+        LLMPlanningAgent(client, roles),
         operators={"don"},
         roles=roles,
-        executor=LLMExecutor(client, store, roles, workspace, max_turns=10),
+        executor=LLMExecutor(client, roles, workspace, max_turns=10),
     )
     orchestrator.start()
 

tests/integration/test_passthrough_live.py

@@ -24,7 +24,7 @@ pytestmark = pytest.mark.integration
 def store_with(tmp_path: Path, *events) -> StateStore:
     store = StateStore(tmp_path / "state.db", oplog=OpLog("nisse-dev", io.StringIO()))
     for event in events:
-        store.append_event(event)
+        store.transcript("t", "helper").append(event)
     return store
 
 
@@ -35,7 +35,7 @@ def test_passthrough_completes_a_plain_turn(
         tmp_path, UserMessage(text="Reply with exactly: pong", source="live")
     )
     result = passthrough.complete(
-        store.read_events(), None, live_model, system="Be terse."
+        store.transcript("t", "helper").read(), None, live_model, system="Be terse."
     )
     texts = [e.text for e in result if isinstance(e, AssistantText)]
     assert texts and "pong" in texts[0].lower()
@@ -53,7 +53,7 @@ def test_passthrough_round_trips_a_tool_call(
     )
     client = passthrough
     produced = client.complete(
-        store.read_events(),
+        store.transcript("t", "helper").read(),
         EXECUTOR_TOOLS,
         live_model,
         system="Use a tool. Do not answer in prose.",
@@ -64,11 +64,13 @@ def test_passthrough_round_trips_a_tool_call(
     # Feed the result back: the follow-up render must be accepted, which is
     # what proves tool_result placement is right for this vendor.
     for event in produced:
-        store.append_event(event)
-    store.append_event(
+        store.transcript("t", "helper").append(event)
+    store.transcript("t", "helper").append(
         ToolResult(call_id=calls[0].call_id, result="app.py\ntests/", success=True)
     )
-    follow_up = client.complete(store.read_events(), EXECUTOR_TOOLS, live_model)
+    follow_up = client.complete(
+        store.transcript("t", "helper").read(), EXECUTOR_TOOLS, live_model
+    )
     assert follow_up, "the vendor rejected or ignored the tool_result turn"
 
 

tests/integration/test_planning_live.py

@@ -27,7 +27,7 @@ def test_live_planning_turn_returns_a_tool_action(
     model = live_model
     roles = RoleMap(executor=model, planner=model, helper=model, expert=model)
     store = StateStore(tmp_path / "state.db", oplog=OpLog("nisse-dev", io.StringIO()))
-    store.append_event(
+    store.transcript("t", "operator").append(
         UserMessage(
             text=(
                 "Project 'hello-cli': a Python script that prints a greeting. "
@@ -36,12 +36,12 @@ def test_live_planning_turn_returns_a_tool_action(
             source="live-test",
         )
     )
-    agent = LLMPlanningAgent(litellm, store, roles)
+    agent = LLMPlanningAgent(litellm, roles)
 
-    action = agent.plan("arch", "hello-cli")
+    action = agent.plan(store.transcript("t", "planner"), "arch", "hello-cli")
 
     assert isinstance(action, SubmitPlan | AskHuman)
     text = action.content if isinstance(action, SubmitPlan) else action.question
     assert text.strip(), "tool call carried empty content"
     # The turn must be durable regardless of what the model said (L8/L7).
-    assert len(store.read_events()) > 1
+    assert len(store.transcript("t", "planner").read()) > 1