Stop sending strict:null in Codex/Responses tool defs

The gpt-account (Codex) turn built its function tools with strict:null. The
Responses API expects strict to be a boolean or omitted; null is malformed,
so the endpoint dropped the whole tool list — leaving the model with nothing
to call. That is exactly the "model returned no tool calls" failure on the
GPT-account path. Omit strict when null; keep the boolean otherwise. Locked
with a test, alongside the Bedrock/Gemini forced-tool-choice encoding.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

AgentCompany-New/src/main/services/provider-tool-turns/responses-format.ts

@@ -7,7 +7,16 @@ import { isRecord, type JsonRecord, type ProviderEventStreamMessage } from "../p
 import { parseArguments, type AgentTurnMessage, type AgentTurnResult, type ToolCall, type ToolDefinition } from "../provider-tool-turn";
 
 export function responsesToolDefs(tools: ToolDefinition[], strict: boolean | null = false): unknown[] {
-  return tools.map((tool) => ({ type: "function", name: tool.name, description: tool.description, parameters: tool.parameters, strict }));
+  // `strict` must be a boolean or omitted. Passing `strict: null` (the Codex
+  // path) is malformed and makes the endpoint drop the whole tool list, so the
+  // model has nothing to call. Omit the field instead of sending null.
+  return tools.map((tool) => ({
+    type: "function",
+    name: tool.name,
+    description: tool.description,
+    parameters: tool.parameters,
+    ...(strict === null ? {} : { strict }),
+  }));
 }
 
 export function toResponsesInput(messages: AgentTurnMessage[]): unknown[] {

AgentCompany-New/tests/unit/provider-tool-turns.test.ts

@@ -4,7 +4,7 @@ import { join } from "node:path";
 import { describe, expect, test } from "vitest";
 import type { JsonRecord, ProviderEventStreamMessage } from "../../src/main/services/provider-http";
 import type { AgentTurnMessage } from "../../src/main/services/provider-tool-turn";
-import { accumulateResponsesStream, parseResponsesOutput, toResponsesInput } from "../../src/main/services/provider-tool-turns/responses-format";
+import { accumulateResponsesStream, parseResponsesOutput, responsesToolDefs, toResponsesInput } from "../../src/main/services/provider-tool-turns/responses-format";
 import { parseChatChoice, toChatMessages } from "../../src/main/services/provider-tool-turns/chat-format";
 import { parseAnthropicContent, toAnthropicMessages } from "../../src/main/services/provider-tool-turns/anthropic-format";
 import { accumulateGeminiStream, antigravityRequestBody } from "../../src/main/services/provider-tool-turns/gemini-format";
@@ -209,3 +209,23 @@ describe("credential resolver", () => {
     })();
   });
 });
+
+describe("tool definitions and forced tool choice", () => {
+  const tools = [{ name: "hire_agent", description: "Add a report.", parameters: { type: "object", properties: {} } }];
+
+  test("Responses tool defs omit strict when null, keep it when boolean", () => {
+    const codex = responsesToolDefs(tools, null)[0] as Record<string, unknown>;
+    expect("strict" in codex).toBe(false); // null would make the endpoint drop the tools
+    expect(responsesToolDefs(tools)[0]).toMatchObject({ type: "function", name: "hire_agent", strict: false });
+    expect(responsesToolDefs(tools, true)[0]).toMatchObject({ strict: true });
+  });
+
+  test("Bedrock and Gemini encode a forced tool choice", () => {
+    expect(bedrockToolConfig(tools, "required").toolChoice).toEqual({ any: {} });
+    expect(bedrockToolConfig(tools).toolChoice).toEqual({ auto: {} });
+    const vertex = vertexRequestBody({ model: "m", instructions: null, messages: [], tools, toolChoice: "required" });
+    expect(vertex.toolConfig).toEqual({ functionCallingConfig: { mode: "ANY" } });
+    const anti = antigravityRequestBody({ model: "m", projectId: "p", instructions: null, messages: [], tools, requestId: "r" });
+    expect((anti.request as Record<string, unknown>).toolConfig).toEqual({ functionCallingConfig: { mode: "AUTO" } });
+  });
+});