fix: canonicalize extra params model lookup keys

Carries forward https://github.com/openclaw/openclaw/pull/44319 by @HenryXiaoYang.

CHANGELOG.md

@@ -32,6 +32,7 @@ Docs: https://docs.openclaw.ai
 - Shared labels: preserve Unicode combining marks and NFC-equivalent accented text in group/channel slug normalization so non-Latin labels no longer lose meaningful characters. Fixes #58932; carries forward #58942 and #58995. Thanks @fengqing-git, @Starhappysh, and @koen666.
 - Channels/Telegram: include probed video width and height when sending regular Telegram videos, so portrait clips render with the correct orientation instead of being stretched by clients. (#18915) Thanks @storyarcade.
 - Docs/Hetzner: clarify that SSH tunnel access requires `AllowTcpForwarding local` before running `ssh -L`, so hardened VPS sshd configs do not block loopback Gateway access. Fixes #54557; carries forward #54564; refs #54954. Thanks @satishkc7, @blackstrype, and @Aftabbs.
+- Agents/model config: resolve per-model extra params through canonical model keys while preserving legacy double-prefixed fallback entries, so provider-prefixed model ids such as `openrouter/auto` keep their configured runtime params. (#44319) Thanks @HenryXiaoYang.
 - Gateway/shutdown: report structured shutdown warnings and HTTP close timeout warnings through `ShutdownResult` while preserving lifecycle hook hardening. Carries forward #41296. Thanks @edenfunf.
 - Control UI: keep Agents Overview and config-form select dropdowns on their configured value after options render while preserving inherited agent model placeholders. Fixes #40352; carries forward #52948. Thanks @xiaoquanidea.
 - Plugins/QA: prebuild the private QA channel runtime before plugin gauntlet source runs so wrapper CPU/RSS measurements are not polluted by private QA dist rebuild work. Thanks @vincentkoc.

src/agents/pi-embedded-runner-extraparams.test.ts

@@ -466,7 +466,7 @@ describe("applyExtraParamsToAgent", () => {
     };
   }
 
-  function buildAnthropicModelConfig(modelKey: string, params: Record<string, unknown>) {
+  function buildModelConfig(modelKey: string, params: Record<string, unknown>) {
     return {
       agents: {
         defaults: {
@@ -817,6 +817,60 @@ describe("applyExtraParamsToAgent", () => {
     expect(payload.parallel_tool_calls).toBe(false);
   });
 
+  it("uses canonical model config keys for provider-prefixed model ids", () => {
+    const payload = runParallelToolCallsPayloadMutationCase({
+      applyProvider: "openrouter",
+      applyModelId: "openrouter/auto",
+      cfg: {
+        agents: {
+          defaults: {
+            models: {
+              "openrouter/auto": {
+                params: {
+                  parallel_tool_calls: false,
+                },
+              },
+            },
+          },
+        },
+      },
+      model: {
+        api: "openai-completions",
+        provider: "openrouter",
+        id: "openrouter/auto",
+      } as Model<"openai-completions">,
+    });
+
+    expect(payload.parallel_tool_calls).toBe(false);
+  });
+
+  it("keeps legacy double-prefixed model config fallback for provider-prefixed model ids", () => {
+    const payload = runParallelToolCallsPayloadMutationCase({
+      applyProvider: "openrouter",
+      applyModelId: "openrouter/auto",
+      cfg: {
+        agents: {
+          defaults: {
+            models: {
+              "openrouter/openrouter/auto": {
+                params: {
+                  parallel_tool_calls: false,
+                },
+              },
+            },
+          },
+        },
+      },
+      model: {
+        api: "openai-completions",
+        provider: "openrouter",
+        id: "openrouter/auto",
+      } as Model<"openai-completions">,
+    });
+
+    expect(payload.parallel_tool_calls).toBe(false);
+  });
+
   it("strips store from proxied openai-completions payloads", () => {
     const payload = runResponsesPayloadMutationCase({
       applyProvider: "google",
@@ -2445,7 +2499,7 @@ describe("applyExtraParamsToAgent", () => {
 
   it("adds Anthropic 1M beta header when context1m is enabled for Opus/Sonnet", () => {
     const { calls, agent } = createOptionsCaptureAgent();
-    const cfg = buildAnthropicModelConfig("anthropic/claude-opus-4-6", { context1m: true });
+    const cfg = buildModelConfig("anthropic/claude-opus-4-6", { context1m: true });
 
     applyExtraParamsToAgent(agent, cfg, "anthropic", "claude-opus-4-6");
 
@@ -2472,7 +2526,7 @@ describe("applyExtraParamsToAgent", () => {
   });
 
   it("does not add Anthropic 1M beta header when context1m is not enabled", () => {
-    const cfg = buildAnthropicModelConfig("anthropic/claude-opus-4-6", {
+    const cfg = buildModelConfig("anthropic/claude-opus-4-6", {
       temperature: 0.2,
     });
     const headers = runAnthropicHeaderCase({
@@ -2529,7 +2583,7 @@ describe("applyExtraParamsToAgent", () => {
   });
 
   it("merges existing anthropic-beta headers with configured betas", () => {
-    const cfg = buildAnthropicModelConfig("anthropic/claude-sonnet-4-5", {
+    const cfg = buildModelConfig("anthropic/claude-sonnet-4-5", {
       context1m: true,
       anthropicBeta: ["files-api-2025-04-14"],
     });
@@ -2549,7 +2603,7 @@ describe("applyExtraParamsToAgent", () => {
   });
 
   it("ignores context1m for non-Opus/Sonnet Anthropic models", () => {
-    const cfg = buildAnthropicModelConfig("anthropic/claude-haiku-3-5", { context1m: true });
+    const cfg = buildModelConfig("anthropic/claude-haiku-3-5", { context1m: true });
     const headers = runAnthropicHeaderCase({
       cfg,
       modelId: "claude-haiku-3-5",

src/agents/pi-embedded-runner/extra-params.ts

@@ -10,6 +10,7 @@ import {
   wrapProviderStreamFn as wrapProviderStreamFnRuntime,
 } from "../../plugins/provider-hook-runtime.js";
 import type { ProviderRuntimeModel } from "../../plugins/provider-runtime-model.types.js";
+import { legacyModelKey, modelKey } from "../model-selection-normalize.js";
 import { supportsGptParallelToolCallsPayload } from "../provider-api-families.js";
 import { resolveProviderRequestPolicyConfig } from "../provider-request-config.js";
 import { createGoogleThinkingPayloadWrapper } from "./google-stream-wrappers.js";
@@ -71,8 +72,11 @@ export function resolveExtraParams(params: {
   agentId?: string;
 }): Record<string, unknown> | undefined {
   const defaultParams = params.cfg?.agents?.defaults?.params ?? undefined;
-  const modelKey = `${params.provider}/${params.modelId}`;
-  const modelConfig = params.cfg?.agents?.defaults?.models?.[modelKey];
+  const canonicalKey = modelKey(params.provider, params.modelId);
+  const legacyKey = legacyModelKey(params.provider, params.modelId);
+  const configuredModels = params.cfg?.agents?.defaults?.models;
+  const modelConfig =
+    configuredModels?.[canonicalKey] ?? (legacyKey ? configuredModels?.[legacyKey] : undefined);
   const globalParams = modelConfig?.params ? { ...modelConfig.params } : undefined;
   const agentParams =
     params.agentId && params.cfg?.agents?.list