fix(task): let task.softRequestBudget lower bundled subagent budgets

The soft request budget resolved to `SOFT_REQUEST_BUDGET[agent.name] ??
configured`, so the bundled entries for scout and sonic replaced the
configured value outright. Lowering `task.softRequestBudget` to tighten
the guard therefore did nothing for exactly the two agents that spawn
most often: a scout kept its 100-request budget no matter how small the
user set the knob. Only 0 (disable) and raising the value for
non-bundled agents had any effect.

Treat both numbers as upper bounds and take the smaller one. The bundled
entries stay ceilings, so a runaway scout is still stopped at 100 by
default and existing behavior is unchanged for anyone who has not
lowered the setting; a configured 0 still disables the guard entirely.
Resolution moves into `resolveSoftRequestBudget`, which also normalizes
negative and fractional inputs, so the rule is testable without standing
up a subprocess run.

This composes with `task.maxEffort` on a separate axis: effort caps how
hard each request thinks, this caps how many requests a run may spend.

(cherry picked from commit f0db29f8f725f11390b64ca9342300c482ff5c5d)

packages/coding-agent/test/task/executor-soft-budget.test.ts

@@ -8,7 +8,7 @@ import { AgentRegistry } from "@oh-my-pi/pi-coding-agent/registry/agent-registry
 import type { CreateAgentSessionResult } from "@oh-my-pi/pi-coding-agent/sdk";
 import * as sdkModule from "@oh-my-pi/pi-coding-agent/sdk";
 import type { AgentSession, AgentSessionEvent, PromptOptions } from "@oh-my-pi/pi-coding-agent/session/agent-session";
-import { runSubprocess } from "@oh-my-pi/pi-coding-agent/task/executor";
+import { resolveSoftRequestBudget, runSubprocess } from "@oh-my-pi/pi-coding-agent/task/executor";
 import type { AgentDefinition } from "@oh-my-pi/pi-coding-agent/task/types";
 import { EventBus } from "@oh-my-pi/pi-coding-agent/utils/event-bus";
 import { TempDir } from "@oh-my-pi/pi-utils";
@@ -105,8 +105,8 @@ function mockCreateAgentSession(session: AgentSession) {
 		eventBus: new EventBus(),
 	} satisfies CreateAgentSessionResult);
 }
-// Named "task": bundled scout/sonic budgets are built-in and override the
-// `task.softRequestBudget` setting, which these tests pin to a tiny value.
+// Named "task": the agent name only selects which bundled ceiling applies, and
+// these tests pin `task.softRequestBudget` to a tiny value below every ceiling.
 const baseAgent: AgentDefinition = {
 	name: "task",
 	description: "test",
@@ -272,3 +272,25 @@ describe("runSubprocess soft request budget", () => {
 		expect(receipt.error).toMatch(new RegExp(`history://${id}`));
 	});
 });
+
+describe("resolveSoftRequestBudget", () => {
+	it("lets a configured budget lower a bundled agent's ceiling", () => {
+		expect(resolveSoftRequestBudget("scout", 20)).toBe(20);
+		expect(resolveSoftRequestBudget("sonic", 20)).toBe(20);
+	});
+
+	it("keeps the bundled ceiling when the configured budget is higher", () => {
+		expect(resolveSoftRequestBudget("scout", 200)).toBe(100);
+		expect(resolveSoftRequestBudget("sonic", 200)).toBe(100);
+	});
+
+	it("uses the configured budget for agents without a bundled entry", () => {
+		expect(resolveSoftRequestBudget("task", 20)).toBe(20);
+	});
+
+	it("keeps 0 disabled and normalizes negative or fractional budgets", () => {
+		expect(resolveSoftRequestBudget("scout", 0)).toBe(0);
+		expect(resolveSoftRequestBudget("scout", -5)).toBe(0);
+		expect(resolveSoftRequestBudget("scout", 20.9)).toBe(20);
+	});
+});