fix(plugins): honor inferred agent model defaults

CHANGELOG.md

@@ -85,7 +85,7 @@ Docs: https://docs.openclaw.ai
 - ACP/OpenCode: update the bundled acpx runtime to 0.6.0 and cover the OpenCode ACP bind path in Docker live tests.
 - Providers/OpenCode Go: add DeepSeek V4 Pro and DeepSeek V4 Flash to the Go catalog while the bundled Pi registry catches up. Fixes #71587.
 - Providers/OpenCode Go: route DeepSeek V4 Pro/Flash through the OpenAI-compatible Go endpoint and suppress invalid `reasoning_effort: "off"` payloads, fixing tool-enabled requests for `opencode-go/deepseek-v4-flash`. Fixes #71683.
-- Plugins/Skill Workshop: run the LLM reviewer on the configured agent default model instead of the hardcoded OpenAI SDK fallback when hook context lacks model metadata. Fixes #71659.
+- Plugins/model defaults: run Skill Workshop review, Active Memory recall, and session-memory slug generation on the configured agent default model instead of the hardcoded OpenAI SDK fallback when hook context lacks model metadata. Fixes #71659.
 - Providers/Venice: fill the required DeepSeek V4 `reasoning_content` placeholder for `venice/deepseek-v4-pro` and `venice/deepseek-v4-flash` replay turns without sending native DeepSeek `thinking` controls that Venice rejects. Fixes #71628.
 - Browser/existing-session: support per-profile Chrome MCP command/args, map `cdpUrl` to `--browserUrl` or `--wsEndpoint`, and avoid combining endpoint flags with `--userDataDir`. Fixes #47879, #48037, and #62706. Thanks @puneet1409, @zhehao, and @madkow1001.
 - Media/plugins: bound MIME sniffing and ZIP archive preflight before handing

extensions/active-memory/index.test.ts

@@ -922,6 +922,53 @@ describe("active-memory plugin", () => {
     });
   });
 
+  it("infers the configured provider for bare active-memory default models", async () => {
+    api.config = {
+      agents: {
+        defaults: {
+          model: { primary: "gpt-5.5" },
+        },
+      },
+      models: {
+        providers: {
+          "openai-codex": {
+            baseUrl: "https://chatgpt.com/backend-api/codex",
+            models: [
+              {
+                id: "gpt-5.5",
+                name: "GPT 5.5",
+                reasoning: true,
+                input: ["text"],
+                cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+                contextWindow: 200_000,
+                maxTokens: 128_000,
+              },
+            ],
+          },
+        },
+      },
+    };
+    api.pluginConfig = {
+      agents: ["main"],
+    };
+    plugin.register(api as unknown as OpenClawPluginApi);
+
+    await hooks.before_prompt_build(
+      { prompt: "what wings should i order? bare model default", messages: [] },
+      {
+        agentId: "main",
+        trigger: "user",
+        sessionKey: "agent:main:main",
+        messageProvider: "webchat",
+      },
+    );
+
+    expect(runEmbeddedPiAgent.mock.calls.at(-1)?.[0]).toMatchObject({
+      provider: "openai-codex",
+      model: "gpt-5.5",
+    });
+  });
+
   it("skips recall when no model or explicit fallback resolves", async () => {
     api.config = {};
     api.pluginConfig = {

extensions/active-memory/index.ts

@@ -7,6 +7,7 @@ import {
   resolveAgentDir,
   resolveAgentEffectiveModelPrimary,
   resolveAgentWorkspaceDir,
+  resolveDefaultModelForAgent,
 } from "openclaw/plugin-sdk/agent-runtime";
 import {
   resolveLivePluginConfigObject,
@@ -1550,13 +1551,11 @@ function extractRecentTurns(messages: unknown[]): ActiveRecallRecentTurn[] {
   return turns;
 }
 
-function parseModelCandidate(modelRef: string | undefined) {
+function parseModelCandidate(modelRef: string | undefined, defaultProvider = DEFAULT_PROVIDER) {
   if (!modelRef) {
     return undefined;
   }
-  return (
-    parseModelRef(modelRef, DEFAULT_PROVIDER) ?? { provider: DEFAULT_PROVIDER, model: modelRef }
-  );
+  return parseModelRef(modelRef, defaultProvider) ?? { provider: defaultProvider, model: modelRef };
 }
 
 function getModelRef(
@@ -1570,14 +1569,20 @@ function getModelRef(
 ): { provider: string; model: string } | undefined {
   const currentRunModel =
     ctx?.modelProviderId && ctx?.modelId ? `${ctx.modelProviderId}/${ctx.modelId}` : undefined;
+  const configuredDefaultModel = resolveAgentEffectiveModelPrimary(api.config, agentId)
+    ? resolveDefaultModelForAgent({ cfg: api.config, agentId })
+    : undefined;
+  const defaultProvider = configuredDefaultModel?.provider ?? DEFAULT_PROVIDER;
   const candidates = [
     config.model,
     currentRunModel,
-    resolveAgentEffectiveModelPrimary(api.config, agentId),
+    configuredDefaultModel
+      ? `${configuredDefaultModel.provider}/${configuredDefaultModel.model}`
+      : undefined,
     config.modelFallback,
   ];
   for (const candidate of candidates) {
-    const parsed = parseModelCandidate(candidate);
+    const parsed = parseModelCandidate(candidate, defaultProvider);
     if (parsed) {
       return parsed;
     }

src/hooks/llm-slug-generator.test.ts

@@ -7,7 +7,13 @@ vi.mock("../agents/agent-scope.js", () => ({
   resolveDefaultAgentId: vi.fn(() => "main"),
   resolveAgentWorkspaceDir: vi.fn(() => "/tmp/openclaw-agent"),
   resolveAgentDir: vi.fn(() => "/tmp/openclaw-agent/.openclaw-agent"),
-  resolveAgentEffectiveModelPrimary: vi.fn(() => null),
+  resolveAgentEffectiveModelPrimary: vi.fn((cfg: OpenClawConfig) => {
+    const model = cfg.agents?.defaults?.model;
+    if (typeof model === "string") {
+      return model;
+    }
+    return model?.primary;
+  }),
 }));
 
 vi.mock("../agents/pi-embedded.js", () => ({
@@ -58,4 +64,43 @@ describe("generateSlugViaLLM", () => {
       }),
     );
   });
+
+  it("infers provider metadata for bare configured agent models", async () => {
+    await generateSlugViaLLM({
+      sessionContent: "hello",
+      cfg: {
+        agents: {
+          defaults: {
+            model: { primary: "gpt-5.5" },
+          },
+        },
+        models: {
+          providers: {
+            "openai-codex": {
+              baseUrl: "https://chatgpt.com/backend-api/codex",
+              models: [
+                {
+                  id: "gpt-5.5",
+                  name: "GPT 5.5",
+                  reasoning: true,
+                  input: ["text"],
+                  cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+                  contextWindow: 200_000,
+                  maxTokens: 128_000,
+                },
+              ],
+            },
+          },
+        },
+      } as OpenClawConfig,
+    });
+
+    expect(runEmbeddedPiAgentMock).toHaveBeenCalledOnce();
+    expect(runEmbeddedPiAgentMock.mock.calls[0]?.[0]).toEqual(
+      expect.objectContaining({
+        provider: "openai-codex",
+        model: "gpt-5.5",
+      }),
+    );
+  });
 });

src/hooks/llm-slug-generator.ts

@@ -9,10 +9,8 @@ import {
   resolveDefaultAgentId,
   resolveAgentWorkspaceDir,
   resolveAgentDir,
-  resolveAgentEffectiveModelPrimary,
 } from "../agents/agent-scope.js";
-import { DEFAULT_PROVIDER, DEFAULT_MODEL } from "../agents/defaults.js";
-import { parseModelRef } from "../agents/model-selection.js";
+import { resolveDefaultModelForAgent } from "../agents/model-selection.js";
 import { runEmbeddedPiAgent } from "../agents/pi-embedded.js";
 import { resolveAgentTimeoutMs } from "../agents/timeout.js";
 import type { OpenClawConfig } from "../config/types.openclaw.js";
@@ -55,11 +53,10 @@ ${params.sessionContent.slice(0, 2000)}
 
 Reply with ONLY the slug, nothing else. Examples: "vendor-pitch", "api-design", "bug-fix"`;
 
-    // Resolve model from agent config instead of using hardcoded defaults
-    const modelRef = resolveAgentEffectiveModelPrimary(params.cfg, agentId);
-    const parsed = modelRef ? parseModelRef(modelRef, DEFAULT_PROVIDER) : null;
-    const provider = parsed?.provider ?? DEFAULT_PROVIDER;
-    const model = parsed?.model ?? DEFAULT_MODEL;
+    const { provider, model } = resolveDefaultModelForAgent({
+      cfg: params.cfg,
+      agentId,
+    });
     const timeoutMs = resolveSlugGeneratorTimeoutMs(params.cfg);
 
     const result = await runEmbeddedPiAgent({