Make routines act on their own and check every bot action

Scheduled runs are pre-approved: they carry out their brief instead of
drafting for confirmation. A run that can't do its brief calls
report_run_failed and is retried 5 and 10 minutes later; only then does
the user get one "failed after N tries" message, and the routine keeps
why in last_error.

Every reply and task now gets ActionCheckNote: look for an existing
record before creating one, and verify an action took effect before
saying it's done. Memory and instructions saves are read back, and
create_bot, create_schedule, start_task and delegate_task refuse
duplicates.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

src/utils/scheduled-run-retry-at.test.ts

@@ -0,0 +1,18 @@
+import { describe, expect, test } from "bun:test";
+import { AppData } from "@/src/data/app-data";
+import scheduledRunRetryAt from "./scheduled-run-retry-at";
+
+describe("scheduledRunRetryAt", () => {
+    test("waits the retry delay after the first failure", () => {
+        expect(scheduledRunRetryAt({ attempts: 1, now: 0 })).toBe(AppData.ScheduledRunRetryMs);
+    });
+
+    test("doubles after each failure", () => {
+        expect(scheduledRunRetryAt({ attempts: 2, now: 0 })).toBe(AppData.ScheduledRunRetryMs * 2);
+        expect(scheduledRunRetryAt({ attempts: 3, now: 0 })).toBe(AppData.ScheduledRunRetryMs * 4);
+    });
+
+    test("treats no attempts as the first", () => {
+        expect(scheduledRunRetryAt({ attempts: 0, now: 10 })).toBe(10 + AppData.ScheduledRunRetryMs);
+    });
+});