test: tighten session_status run-session proof (#82696)

CHANGELOG.md

@@ -28,6 +28,7 @@ Docs: https://docs.openclaw.ai
 - Providers/OpenAI Codex: include base `gpt-5.5` and `gpt-5.4` reasoning metadata in the bundled Codex catalog so `/think xhigh` remains available for those models. Fixes #82744.
 - Agents/edit tool: honor `file_path` and related path aliases when resolving edit-recovery targets, so post-write errors no longer surface false edit failures after the file actually changed. Fixes #81909. Thanks @giodl73-repo.
 - QQBot: treat only explicit truthy `QQBOT_DEBUG` values as enabling debug logs, so false-like values such as `0` no longer expose debug output. Fixes #82644. (#82697) Thanks @leno23.
+- Agents/session_status: resolve implicit no-arg status lookups against the live run session, so `/think` changes report the current thinking level instead of stale sandbox state. Fixes #82669. (#82696) Thanks @leno23.
 - Gateway/diagnostics: add opt-in critical memory pressure stability snapshots with gateway logs, V8 heap, cgroup, active-resource, and redacted large session-file evidence. Fixes #82518.
 - Doctor/Gateway: avoid treating unrelated macOS LaunchAgents as legacy gateways just because their environment values mention old checkout paths.
 - CLI/setup: collapse raw gateway config keys in existing-config summaries into friendly `Model` and `Gateway` rows.

scripts/repro/session-status-run-session-key-live-proof.mjs

@@ -6,9 +6,9 @@
 import fs from "node:fs";
 import os from "node:os";
 import path from "node:path";
-import { createSessionStatusTool } from "../../src/agents/tools/session-status-tool.ts";
 
 const tmpRoot = fs.mkdtempSync(path.join(os.tmpdir(), "openclaw-session-status-proof-"));
+const configPath = path.join(tmpRoot, "openclaw.json");
 const storePath = path.join(tmpRoot, "sessions.json");
 const store = {
   "agent:main:telegram:default:direct:1234": {
@@ -24,13 +24,32 @@ const store = {
     thinkingLevel: "high",
   },
 };
+fs.writeFileSync(configPath, "{}\n");
 fs.writeFileSync(storePath, `${JSON.stringify(store, null, 2)}\n`);
+process.env.OPENCLAW_CONFIG_PATH = configPath;
+process.env.OPENCLAW_DISABLE_BUNDLED_PLUGINS = "1";
+
+const originalStderrWrite = process.stderr.write.bind(process.stderr);
+process.stderr.write = (chunk, encoding, callback) => {
+  const text = String(chunk);
+  if (text.includes("gateway connect failed:")) {
+    if (typeof encoding === "function") {
+      encoding();
+    } else if (typeof callback === "function") {
+      callback();
+    }
+    return true;
+  }
+  return originalStderrWrite(chunk, encoding, callback);
+};
+
+const { createSessionStatusTool } = await import("../../src/agents/tools/session-status-tool.ts");
 
 const config = {
   session: { mainKey: "main", scope: "per-sender", store: storePath },
   agents: {
     defaults: {
-      model: { primary: "openai/gpt-5.4" },
+      model: { primary: "proof/gpt-5.4" },
       models: {},
     },
   },
@@ -39,22 +58,37 @@ const config = {
   },
 };
 
-const tool = createSessionStatusTool({
-  agentSessionKey: "agent:main:telegram:default:direct:1234",
-  runSessionKey: "agent:main:main",
-  config,
-});
+try {
+  const tool = createSessionStatusTool({
+    agentSessionKey: "agent:main:telegram:default:direct:1234",
+    runSessionKey: "agent:main:main",
+    config,
+  });
 
-const result = await tool.execute("live-proof-implicit-run-session", {});
-const text = typeof result === "string" ? result : JSON.stringify(result);
-const thinkingMatch = text.match(/think(?:ing)?[:\s]+(\w+)/i);
+  const result = await tool.execute("live-proof-implicit-run-session", {});
+  const text =
+    typeof result === "string"
+      ? result
+      : (result.content.find((item) => item.type === "text")?.text ?? result.details.statusText);
+  const thinkingMatch = text.match(/\bThink:\s+(\w+)/i);
+  const sessionKey = typeof result === "string" ? undefined : result.details.sessionKey;
 
-console.log(
-  "implicit session_status resolved thinkingLevel from store =",
-  store["agent:main:main"].thinkingLevel,
-);
-console.log("status text mentions thinking:", thinkingMatch?.[1] ?? "(see full status below)");
-console.log("--- status excerpt ---");
-console.log(text.split("\n").slice(0, 12).join("\n"));
+  if (sessionKey !== "agent:main:main") {
+    throw new Error(`expected details.sessionKey agent:main:main, got ${String(sessionKey)}`);
+  }
+  if (thinkingMatch?.[1] !== "high") {
+    throw new Error(`expected status text to mention Think: high, got ${thinkingMatch?.[1]}`);
+  }
 
-fs.rmSync(tmpRoot, { recursive: true, force: true });
+  console.log(
+    "implicit session_status resolved thinkingLevel from store =",
+    store["agent:main:main"].thinkingLevel,
+  );
+  console.log("status text mentions thinking:", thinkingMatch[1]);
+  console.log("details.sessionKey =", sessionKey);
+  console.log("--- status excerpt ---");
+  console.log(text.split("\n").slice(0, 8).join("\n"));
+} finally {
+  process.stderr.write = originalStderrWrite;
+  fs.rmSync(tmpRoot, { recursive: true, force: true });
+}

src/agents/openclaw-tools.session-status.test.ts

@@ -556,7 +556,10 @@ describe("session_status tool", () => {
       config: mockConfig as never,
     });
 
-    await tool.execute("call-implicit-run-session-thinking", {});
+    const result = await tool.execute("call-implicit-run-session-thinking", {});
+    const details = result.details as { ok?: boolean; sessionKey?: string };
+    expect(details.ok).toBe(true);
+    expect(details.sessionKey).toBe("agent:main:main");
 
     const statusArg = mockCallArg(buildStatusMessageMock) as Record<string, unknown>;
     const sessionEntry = statusArg.sessionEntry as SessionEntry;