Classify thrown context failures

Author: gitcommit90

package-lock.json

@@ -1,7 +1,7 @@
 {
   "name": "@gitcommit90/rerouted",
   "productName": "ReRouted",
-  "version": "0.5.3",
+  "version": "0.5.4",
   "description": "A local AI router for connected accounts, models, named routes, and automatic fallback.",
   "author": "gitcommit90",
   "license": "MIT",

package.json

@@ -1110,13 +1110,15 @@ function createRouter({ store, fetchImpl = fetch, requestLog, timeoutMs, usage,
           defaultCooldownMs: COOLDOWN_MS.transient,
         };
       }
+      const error = e.message || String(e);
+      const classification = classifyFailure(502, error);
       return {
         ok: false,
-        status: 502,
-        error: e.message || String(e),
-        cooldownEligible: true,
-        failureKind: "transient",
-        defaultCooldownMs: COOLDOWN_MS.transient,
+        status: classification.kind === "request" ? 400 : 502,
+        error,
+        cooldownEligible: classification.eligible,
+        failureKind: classification.kind,
+        defaultCooldownMs: classification.defaultCooldownMs,
       };
     } finally {
       if (!cleanupDeferred) bound.cleanup();

src/lib/router.js

@@ -199,6 +199,42 @@ describe("same-provider OAuth account fallback", () => {
     assert.equal(logger.entries.find((entry) => entry.meta?.event === "accounts_exhausted").meta.status, 400);
   });
 
+  it("does not lock accounts when a context failure is thrown while preparing a stream", async () => {
+    const store = createStore(tmpConfig());
+    store.seed({
+      providers: [
+        chatgptAccount("prov_a", "token-a", 100),
+        chatgptAccount("prov_b", "token-b", 200),
+      ],
+    });
+    const calls = [];
+    const router = createRouter({
+      store,
+      logger: captureLogger(),
+      fetchImpl: async (_url, options) => {
+        calls.push(authToken(options));
+        throw new Error(
+          "Your input exceeds the context window of this model. Please adjust your input and try again."
+        );
+      },
+    });
+
+    const result = await router.chatCompletions({
+      body: {
+        model: "chatgpt/gpt-5.4",
+        messages: [{ role: "user", content: "oversized request" }],
+        stream: true,
+      },
+    });
+
+    assert.equal(result.ok, false);
+    assert.equal(result.status, 400);
+    assert.deepEqual(calls, ["token-a", "token-b"]);
+    for (const provider of store.load().providers) {
+      assert.equal(provider.modelLocks?.["gpt-5.4"], undefined);
+    }
+  });
+
   it("assigns monotonic oauth aliases and advertises only shared model ids", () => {
     const configPath = tmpConfig();
     const store = createStore(configPath);