fix: allow subagent thinking config patch (#75802)

CHANGELOG.md

@@ -9,6 +9,7 @@ Docs: https://docs.openclaw.ai
 - Agents/Codex: stop prompting message-tool-only source turns to finish with `NO_REPLY`, so quiet turns are represented by not calling the visible message tool instead of conflicting final-text instructions. Thanks @pashpashpash.
 - Gateway/config: report failed backup restores as failed in logs and config observe audit records instead of marking them valid. (#70515) Thanks @davidangularme.
 - Compaction: use the active session model fallback chain for implicit summarization failures without persisting fallback model selection, so Azure content-filter 400s can recover. Fixes #64960. (#74470) Thanks @jalehman and @OpenCodeEngineer.
+- Gateway/config: allow `gateway config.patch` to update documented subagent thinking defaults. Fixes #75764. (#75802) Thanks @kAIborg24.
 
 ## 2026.4.30
 

src/agents/tools/gateway-tool-guard-coverage.test.ts

@@ -62,8 +62,10 @@ describe("gateway config mutation guard coverage", () => {
       expect.arrayContaining([
         "agents.defaults.systemPromptOverride",
         "agents.defaults.model",
+        "agents.defaults.subagents.thinking",
         "agents.list[].id",
         "agents.list[].model",
+        "agents.list[].subagents.thinking",
         "channels.*.requireMention",
         "messages.visibleReplies",
         "messages.groupChat.visibleReplies",
@@ -71,6 +73,77 @@ describe("gateway config mutation guard coverage", () => {
     );
   });
 
+  it("allows documented subagent thinking default edits via config.patch", () => {
+    expectAllowed(
+      {},
+      {
+        agents: {
+          defaults: {
+            subagents: { thinking: "medium" },
+          },
+        },
+      },
+    );
+    expectAllowed(
+      {
+        agents: {
+          defaults: {
+            subagents: { thinking: "low" },
+          },
+        },
+      },
+      {
+        agents: {
+          defaults: {
+            subagents: { thinking: "high" },
+          },
+        },
+      },
+    );
+  });
+
+  it("allows documented per-agent subagent thinking edits via config.patch", () => {
+    expectAllowed(
+      {
+        agents: {
+          list: [{ id: "worker", subagents: { thinking: "low" } }],
+        },
+      },
+      {
+        agents: {
+          list: [{ id: "worker", subagents: { thinking: "medium" } }],
+        },
+      },
+    );
+    expectAllowed(
+      { agents: { list: [] as Array<Record<string, unknown>> } },
+      {
+        agents: {
+          list: [{ id: "helper", subagents: { thinking: "medium" } }],
+        },
+      },
+    );
+  });
+
+  it("keeps neighboring subagent policy fields protected via config.patch", () => {
+    expectBlocked(
+      { agents: { defaults: { subagents: { allowAgents: ["worker"] } } } },
+      { agents: { defaults: { subagents: { allowAgents: ["*"] } } } },
+    );
+    expectBlocked(
+      {
+        agents: {
+          list: [{ id: "worker", subagents: { requireAgentId: true } }],
+        },
+      },
+      {
+        agents: {
+          list: [{ id: "worker", subagents: { requireAgentId: false } }],
+        },
+      },
+    );
+  });
+
   it("allows visible reply delivery mode edits via config.patch", () => {
     expectAllowed(
       {},

src/agents/tools/gateway-tool.ts

@@ -34,12 +34,14 @@ const ALLOWED_GATEWAY_CONFIG_PATHS = [
   "agents.defaults.promptOverlays",
   "agents.defaults.model",
   "agents.defaults.thinkingDefault",
+  "agents.defaults.subagents.thinking",
   "agents.defaults.reasoningDefault",
   "agents.defaults.fastModeDefault",
   "agents.list[].id",
   "agents.list[].systemPromptOverride",
   "agents.list[].model",
   "agents.list[].thinkingDefault",
+  "agents.list[].subagents.thinking",
   "agents.list[].reasoningDefault",
   "agents.list[].fastModeDefault",
   // Mention gating is an agent-facing scope knob across channel adapters.