fix(release): resolve inherited base-reds surfaced by v3.8.35 release CI

Cycle base-reds that only run on PR→main (not the PR→release fast-path):

- test(autoCombo): suffixComposition-4517 used node:test in a vitest-only dir
  (#4753) → vitest found no suite. Switch to the vitest API. (Vitest job)
- test(agentSkills): openapiParser fixture wrote docs/reference/openapi.yaml;
  parser reads docs/openapi.yaml since #4781 → point fixture at the new path.
  (Unit/Coverage/Node24/Node26 shard 4)
- test(integration): proxy-pipeline source-scan expected inline streaming-cost
  code that #4790/#3501 extracted to the recordStreamingCost leaf → assert the
  delegation instead. (Integration 1/2)
- fix(chatCore): derive the log trace id from crypto, not Math.random
  (CodeQL js/insecure-randomness — log-correlation id, not a secret).
- test(resilience): circuit-breaker invalid-cooldown fallback asserted t>29000,
  flaking on slow CI where ~1.6s elapsed gave t=28401 → tolerate wall-clock
  drift (t>25000). (Unit 6/8)

open-sse/handlers/chatCore.ts

@@ -353,8 +353,10 @@ export async function handleChatCore({
   const isModelScope = () => isModelScopeProvider(provider, credentials?.providerSpecificData);
   const startTime = Date.now();
   // Per-request trace id + checkpoint helper. Lets us see exactly which await
-  // a hung request was sitting on in `[STAGE_TRACE]` log lines.
-  const traceId = Math.random().toString(36).slice(2, 8);
+  // a hung request was sitting on in `[STAGE_TRACE]` log lines. Uses crypto RNG
+  // (not Math.random) purely to satisfy CodeQL js/insecure-randomness — this id
+  // is a log-correlation token, not a security secret.
+  const traceId = globalThis.crypto.randomUUID().slice(0, 6);
 
   // Emit request.started event for real-time dashboard
   setImmediate(() => {
@@ -1126,20 +1128,21 @@ export async function handleChatCore({
         }
       }
       // Phase 4A: unified output styles (supersedes cavemanOutputMode via the back-compat shim).
-      let outputStyleResult: import("../services/compression/outputStyles/apply.ts").OutputStylesResult | null =
-        null;
+      let outputStyleResult:
+        | import("../services/compression/outputStyles/apply.ts").OutputStylesResult
+        | null = null;
       if (config.enabled) {
         try {
-          const { resolveOutputStyleSelection } = await import(
-            "../services/compression/outputStyles/backCompat.ts"
-          );
+          const { resolveOutputStyleSelection } =
+            await import("../services/compression/outputStyles/backCompat.ts");
           const selection = resolveOutputStyleSelection(config);
           if (selection.length > 0) {
-            const { applyOutputStyles } = await import(
-              "../services/compression/outputStyles/apply.ts"
-            );
+            const { applyOutputStyles } =
+              await import("../services/compression/outputStyles/apply.ts");
             const outputStyleLanguage =
-              config.languageConfig?.enabled === true ? config.languageConfig.defaultLanguage : "en";
+              config.languageConfig?.enabled === true
+                ? config.languageConfig.defaultLanguage
+                : "en";
             outputStyleResult = applyOutputStyles(
               body as Parameters<typeof applyOutputStyles>[0],
               selection,
@@ -1156,7 +1159,10 @@ export async function handleChatCore({
               outputStyleResult.skippedReason &&
               outputStyleResult.skippedReason !== "no_styles"
             ) {
-              log?.debug?.("COMPRESSION", `Output styles skipped: ${outputStyleResult.skippedReason}`);
+              log?.debug?.(
+                "COMPRESSION",
+                `Output styles skipped: ${outputStyleResult.skippedReason}`
+              );
             }
           }
         } catch (err) {
@@ -1175,7 +1181,9 @@ export async function handleChatCore({
         typeof (compressionInputBody as Record<string, unknown>)?.max_tokens === "number"
           ? ((compressionInputBody as Record<string, unknown>).max_tokens as number)
           : null;
-      let adaptiveTelemetry: import("../services/compression/adaptiveCompression/types.ts").AdaptiveTelemetry | null = null;
+      let adaptiveTelemetry:
+        | import("../services/compression/adaptiveCompression/types.ts").AdaptiveTelemetry
+        | null = null;
       const compressionPlan = selectCompressionPlan(
         config,
         compressionComboKey,
@@ -1187,7 +1195,9 @@ export async function handleChatCore({
         {
           modelContextLimit: adaptiveModelContextLimit,
           requestMaxTokens: requestMaxTokens,
-          onAdaptive: (t) => { adaptiveTelemetry = t; },
+          onAdaptive: (t) => {
+            adaptiveTelemetry = t;
+          },
         }
       );
       const mode = compressionPlan.mode as CompressionConfig["defaultMode"];
@@ -1400,12 +1410,10 @@ export async function handleChatCore({
       if (outputStyleResult) {
         void (async () => {
           try {
-            const { buildOutputStyleTelemetry } = await import(
-              "../services/compression/outputStyles/telemetry.ts"
-            );
-            const { insertCompressionRunTelemetryRow } = await import(
-              "../../src/lib/db/compressionRunTelemetry.ts"
-            );
+            const { buildOutputStyleTelemetry } =
+              await import("../services/compression/outputStyles/telemetry.ts");
+            const { insertCompressionRunTelemetryRow } =
+              await import("../../src/lib/db/compressionRunTelemetry.ts");
             const record = buildOutputStyleTelemetry({
               requestId: skillRequestId ?? traceId ?? "",
               model: effectiveModel ?? "",

tests/integration/proxy-pipeline.test.ts

@@ -71,8 +71,11 @@ describe("Chat Pipeline — handleSingleModelChat decomposition", () => {
   });
 
   it("chatCore should record cost for both non-streaming and streaming responses", () => {
+    // Non-streaming cost is still recorded inline; streaming cost was extracted to
+    // the recordStreamingCost leaf (open-sse/handlers/chatCore/streamingCost.ts,
+    // #4790 / #3501), so chatCore now delegates streaming cost to it.
     assert.match(coreSrc, /if \(apiKeyInfo\?\.id && estimatedCost > 0\)/);
-    assert.match(coreSrc, /if \(apiKeyInfo\?\.id && streamUsage\)/);
+    assert.match(coreSrc, /recordStreamingCost\(/);
   });
 });
 
@@ -208,12 +211,22 @@ describe("Plugin Architecture — plugins/hooks.ts", () => {
 
   it("should run onRequest hooks in priority order", async () => {
     const order = [];
-    hooks.registerHook("onRequest", "first", () => {
-      order.push("first");
-    }, 1);
-    hooks.registerHook("onRequest", "second", () => {
-      order.push("second");
-    }, 2);
+    hooks.registerHook(
+      "onRequest",
+      "first",
+      () => {
+        order.push("first");
+      },
+      1
+    );
+    hooks.registerHook(
+      "onRequest",
+      "second",
+      () => {
+        order.push("second");
+      },
+      2
+    );
 
     const ctx = { requestId: "r1", body: {}, model: "test", metadata: {} };
     await hooks.runOnRequest(ctx);
@@ -221,13 +234,23 @@ describe("Plugin Architecture — plugins/hooks.ts", () => {
   });
 
   it("should support request blocking via emitHookBlocking", async () => {
-    hooks.registerHook("onRequest", "blocker", () => ({
-      blocked: true,
-      response: { error: "denied" },
-    }), 1);
-    hooks.registerHook("onRequest", "never-runs", () => {
-      throw new Error("should not run");
-    }, 2);
+    hooks.registerHook(
+      "onRequest",
+      "blocker",
+      () => ({
+        blocked: true,
+        response: { error: "denied" },
+      }),
+      1
+    );
+    hooks.registerHook(
+      "onRequest",
+      "never-runs",
+      () => {
+        throw new Error("should not run");
+      },
+      2
+    );
 
     const ctx = { requestId: "r2", body: {}, model: "test", metadata: {} };
     const result = await hooks.emitHookBlocking("onRequest", ctx);

tests/unit/agentSkills-openapiParser.test.ts

@@ -5,7 +5,8 @@ import path from "node:path";
 import os from "node:os";
 
 // Dynamic import to pick up ESM module
-const { parseOpenapi, getEndpointsForArea } = await import("../../src/lib/agentSkills/openapiParser.ts");
+const { parseOpenapi, getEndpointsForArea } =
+  await import("../../src/lib/agentSkills/openapiParser.ts");
 
 // ─── Fixture helpers ──────────────────────────────────────────────────────────
 
@@ -15,7 +16,9 @@ const { parseOpenapi, getEndpointsForArea } = await import("../../src/lib/agentS
  */
 function withFixtureOpenapi(yamlContent: string): { cleanup: () => void } {
   const tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), "omni-openapi-test-"));
-  const docsDir = path.join(tmpDir, "docs", "reference");
+  // openapiParser reads docs/openapi.yaml (consolidated location since #4781,
+  // previously docs/reference/openapi.yaml) — the fixture must mirror that path.
+  const docsDir = path.join(tmpDir, "docs");
   fs.mkdirSync(docsDir, { recursive: true });
   fs.writeFileSync(path.join(docsDir, "openapi.yaml"), yamlContent, "utf-8");
 
@@ -118,7 +121,7 @@ test("parseOpenapi() groups /api/providers/* under 'providers' area", () => {
     assert.ok(providerOps, "Expected 'providers' area to exist");
     assert.ok(
       providerOps!.length >= 5,
-      `Expected at least 5 provider endpoints, got ${providerOps!.length}`,
+      `Expected at least 5 provider endpoints, got ${providerOps!.length}`
     );
 
     const paths = providerOps!.map((op) => op.path);
@@ -150,11 +153,11 @@ test("parseOpenapi() groups /api/v1/* under 'inference' area", () => {
     assert.ok(inferenceOps, "Expected 'inference' area to exist");
     assert.ok(
       inferenceOps!.length >= 1,
-      `Expected at least 1 inference endpoint, got ${inferenceOps!.length}`,
+      `Expected at least 1 inference endpoint, got ${inferenceOps!.length}`
     );
     assert.ok(
       inferenceOps!.some((op) => op.path === "/api/v1/chat/completions"),
-      "Expected /api/v1/chat/completions in inference area",
+      "Expected /api/v1/chat/completions in inference area"
     );
   } finally {
     cleanup();
@@ -184,7 +187,7 @@ test("parseOpenapi() throws if openapi.yaml is missing", () => {
     assert.throws(
       () => parseOpenapi(),
       /openapiParser: could not read/,
-      "Expected error when openapi.yaml is missing",
+      "Expected error when openapi.yaml is missing"
     );
   } finally {
     process.chdir(originalCwd);
@@ -225,9 +228,9 @@ test(
     assert.ok(providerOps, "Expected 'providers' area in real OpenAPI spec");
     assert.ok(
       providerOps!.length >= 5,
-      `Expected ≥5 provider endpoints in real spec, got ${providerOps!.length}`,
+      `Expected ≥5 provider endpoints in real spec, got ${providerOps!.length}`
     );
-  },
+  }
 );
 
 test(
@@ -237,11 +240,11 @@ test(
     const endpoints = getEndpointsForArea("providers");
     assert.ok(
       endpoints.length >= 5,
-      `Expected ≥5 provider endpoint strings, got ${endpoints.length}: ${endpoints.join(", ")}`,
+      `Expected ≥5 provider endpoint strings, got ${endpoints.length}: ${endpoints.join(", ")}`
     );
     // Each entry should match "METHOD /path"
     for (const ep of endpoints) {
       assert.match(ep, /^[A-Z]+ \//, `Endpoint "${ep}" does not match METHOD /path format`);
     }
-  },
+  }
 );

tests/unit/autoCombo/suffixComposition-4517.test.ts

@@ -10,7 +10,11 @@
  *
  * #4517.
  */
-import { describe, it, before, after, beforeEach } from "node:test";
+// NOTE: tests/unit/autoCombo/ is a vitest-only scope (the node:test runner in
+// test:unit:ci does not walk this dir), so this suite uses the vitest API even
+// though it asserts via node:assert. (#4753 originally landed it with node:test
+// imports → vitest reported "No test suite found" and the test ran nowhere.)
+import { describe, it, beforeAll, afterAll, beforeEach } from "vitest";
 import assert from "node:assert/strict";
 import {
   buildAutoCandidateFilter,
@@ -20,7 +24,7 @@ import {
 describe("suffixComposition :free tier (#4517)", () => {
   const ORIGINAL_ENV = { ...process.env };
 
-  before(() => {
+  beforeAll(() => {
     // Snapshot env so we can restore it after each test.
   });
 
@@ -30,7 +34,7 @@ describe("suffixComposition :free tier (#4517)", () => {
     process.env = { ...ORIGINAL_ENV };
   });
 
-  after(() => {
+  afterAll(() => {
     process.env = ORIGINAL_ENV;
   });
 

tests/unit/circuit-breaker-failure-kind.test.ts

@@ -219,7 +219,11 @@ test("invalid cooldownByKind values (NaN, Infinity, negative) fall back to reset
     cb.reset();
     cb._onFailure(kind);
     const t = cb._timeUntilReset();
-    assert.ok(t > 29_000 && t <= 30_000, `expected resetTimeout fallback for ${kind}, got ${t}`);
+    // _timeUntilReset() = resetTimeout - elapsed-since-failure, so the lower bound
+    // must tolerate real wall-clock drift on slow/loaded CI runners (this loop took
+    // ~1.6s once → t=28401, flaking the old `> 29_000`). Any t well above the invalid
+    // cooldown values (NaN/Infinity/negative) and <= resetTimeout proves the fallback.
+    assert.ok(t > 25_000 && t <= 30_000, `expected resetTimeout fallback for ${kind}, got ${t}`);
   }
   cb.reset();
 });