Implement A9 state-change logging in the state store

Retrofit for the conventions merged in #4 (issue #3):

- OpLog: JSON-lines emitter with the fixed envelope (ts, agent,
  event, task_id, cause); task_id set by the orchestrator at
  unit-of-work start
- StateStore owns emission: transition, intent begin/commit/abort,
  and schema migrations each emit their event right after the
  durable write commits; callers pass the cause
- reconcile_on_boot emits boot.reconcile with action and verdicts,
  and resolves intents with reconciliation-verdict causes
- Failed resolutions emit nothing (a line exists iff the write
  committed)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

nisse-dev/src/nisse_dev/oplog.py

@@ -0,0 +1,33 @@
+"""Structured operational logging (spec A9).
+
+One JSON object per line to stderr (journald captures it when running as a
+service). Every event carries the fixed envelope: ts, agent, event, task_id,
+cause. Logs are diagnostic, never load-bearing — nothing may read them back
+to derive state.
+"""
+
+import json
+import sys
+from datetime import UTC, datetime
+from typing import Any, TextIO
+
+
+class OpLog:
+    def __init__(self, agent: str, stream: TextIO | None = None) -> None:
+        self.agent = agent
+        # Correlation ID for the enclosing unit of work; the orchestrator sets
+        # it when a unit starts and clears it when the unit ends (A9).
+        self.task_id: str | None = None
+        self._stream = stream if stream is not None else sys.stderr
+
+    def emit(self, event: str, *, cause: str, **fields: Any) -> None:
+        record: dict[str, Any] = {
+            "ts": datetime.now(UTC).isoformat(),
+            "agent": self.agent,
+            "event": event,
+            "task_id": self.task_id,
+            "cause": cause,
+            **fields,
+        }
+        self._stream.write(json.dumps(record) + "\n")
+        self._stream.flush()