@rabbi696/ezrouter 0.6.7 → 0.6.9

package/app/CHANGELOG.md

@@ -1,3 +1,83 @@
+# v0.6.9 (2026-09-20)
+
+## Upstream Synchronization & Core Enhancements (decolua/9router #4063 - #4176)
+- **AWS Bedrock Provider**:
+  - Added full AWS Bedrock provider support with AWS SSO and SigV4 request signing (`#4157`).
+  - Added `@aws-sdk/credential-providers` support and extracted shared AWS event stream utilities.
+- **Provider Capabilities & Models**:
+  - Added DeepSeek-V4.1-Flash for Codebuddy-Intl and Ollama (`#4063`).
+  - Added `ainetcafe` provider integration.
+  - Derived Ollama monthly free-plan quota reset date accurately from user signup date (`#4150`).
+  - Forwarded `reasoning_effort` and thinking level overrides to upstream Qoder models via modular protocol pipeline (`#4154`).
+  - Completed HuggingFace Inference Providers router migration (`#4158`).
+  - Configured priority service tier and generalized fast mode pricing for Codex models (`#4166`).
+  - Supported GitHub Copilot automatic model selection (`#4169`).
+  - Exposed context window and max token limits directly on combo `/v1/models` responses (`#4174`).
+  - Routed Union Alpha provider calls through Anthropic Messages API (`#4068`).
+  - Added missing long-context pricing tiers for 128k+ models (`#4149`).
+  - Restricted o-series capabilities pattern matching to avoid false positives on Cline's Solar Pro 4 (`#4141`).
+- **Translators & Schema Enhancements**:
+  - Preserved tool-result images, restored Kiro tool names with underscores, and used neutral placeholder turns for tool-result-only user turns (`#4074`).
+  - Preserved optional tool parameters across Responses and Codex translation layers (`#4142`).
+  - Repaired missing or corrupted `tool_call_id` dropped by Responses clients (`#4145`).
+  - Allowed unkeyed tool outputs in `stripOrphanedToolOutputs` to be properly paired (`#4146`).
+  - Sanitized unsupported JSON schema constraints (`pattern`, `format`, `minItems`) and normalized property definitions for Gemini (`#4147`).
+  - Preserved and normalized `usage.cost.total` for OpenClaw clients (`#4143`).
+- **OAuth, Sessions & Stream Stability**:
+  - Hardened Zed OAuth token refresh lifecycle, added live model support, and supported hosted OAuth callback flow (`#4075`, `#4092`, `#4175`).
+  - Resolved OpenCode 403 `FreeTierError` with canonical session headers and reused stable upstream sessions to stop 429 rate limiting (`#4069`, `#4071`, `#4084`).
+  - Reported client aborts occurring after HTTP 200 in-band to clients instead of silently dropping connections (`#4065`).
+  - Retried CommandCode on transient stream errors and eliminated duplicate stop chunks (`#4138`).
+  - Added `"cache-affinity"` routing strategy to pin conversations to single provider accounts and maximize prompt caching (`#4168`).
+  - Isolated video generation poll cooldowns and enforced pinned account execution (`#4171`).
+  - Guaranteed synced model catalog updates reach live request handlers, keyed by gateway (`#4167`).
+  - Sanitized Hermes system prompts and identity leaks for Antigravity Claude envelope (`#4076`).
+  - Scoped cached Gemini thought signatures to model families in Antigravity (`#4139`).
+  - Stripped Claude Code billing headers from system prompts in Antigravity (`#4137`).
+- **Dashboard & Connection Testing**:
+  - Resolved connection test targets dynamically from provider registry `validateUrl` (`#4159`).
+  - Added opt-in dashboard setting to inspect and log sanitized request payloads (`#4162`).
+  - Fixed `fastModeUpdatingId` state declaration in `ProviderLimits` component.
+
+# v0.6.8 (2026-09-14)
+
+## Upstream Synchronization & Core Enhancements (decolua/9router #3955 - #4062)
+- **Security & Logging Hardening**:
+  - Masked sensitive credential headers in request logs (`#4022`).
+  - Guaranteed MITM hosts cleanup on process shutdown/crash and guarded tool DNS restore (`#4014`).
+- **OAuth & Account Reliability**:
+  - Parsed numeric epoch `expiresAt` timestamps so imported connections reliably refresh (`#3997`).
+  - Prevented account cooldowns for request-scoped 4xx client errors (`#4059`).
+  - Surfaced actionable failure reasons when connection testing fails instead of a generic error (`#4030`).
+  - Reset credential form state on modal close so reopened dialogs start empty (`#4027`).
+- **Streaming & Translation Accuracy**:
+  - Stopped blocking Ollama upstream NDJSON stream as non-SSE body (`#4002`).
+  - Skipped RTK compression/ponytail injection on media-carrying requests (`#4056`).
+  - Stripped `output_config.format` for Claude passthrough targets (`#4050`).
+  - Preserved Responses tool-output images as native image blocks (`#4058`).
+  - Surfaced prompt-cache reads to chat/completions clients and stopped double-counting in usage logs (`#3984`).
+- **Antigravity & Gemini**:
+  - Enforced `maxOutputTokens > thinkingBudget` and properly injected `thinkingConfig` (`#3981`).
+  - Resolved schema map vs schema recursion bug (`#3999`).
+  - Preserved multiple system messages in Gemini `systemInstruction` (`#3973`).
+  - Supported `IMAGE` and `DOCUMENT` blocks for Antigravity Claude envelope (`#3968`).
+- **Provider Capabilities & Models**:
+  - Enabled vision capability on DeepSeek V4.1 (`#4053`).
+  - Declared `thinkingCanDisable: false` on OpenAI reasoning models to clamp to minimal reasoning instead of unsupported `"none"` (`#4048`).
+  - Merged member capabilities, context window, and max completion tokens into combo `/v1/models` entries, flattening nested combos (`#4034`).
+  - Honored operator-declared custom model capabilities over default patterns (`#3974`).
+  - Forwarded CommandCode images instead of `[image omitted]` and formatted credit amounts with dot decimal (`#4004`, `#4025`).
+  - Routed OpenCode Go Muse Spark and Grok models to Responses endpoint with reasoning cleanups and free tool choice normalization (`#3975`, `#4061`, `#4062`).
+  - Added Cline free-tier models and API-key auth support (`#3967`).
+  - Aligned Vertex GA Veo model IDs and resolved polling from self-identifying job IDs (`#3964`, `#3965`).
+  - Improved DeepSeek credit balance display with distinct styling and currency badge (`#3955`, `#4021`).
+  - Configured and validated Google PSE search engine IDs (`#4047`).
+- **CLI & Dashboard UI**:
+  - Persisted model IDs when creating TUI combos (`#4046`).
+  - Added modal fullscreen toggle and persisted quota tracker filter preferences (`#4057`).
+  - Detected Python 3.14 installs on Windows for Headroom extras status (`#3978`).
+  - Integrated Persian (fa) translations and translated session affinity and scheduling UI across all 34 locales (`#4028`, `#4055`).
+
 # v0.6.7 (2026-09-11)
 
 ## Upstream Synchronization & Core Enhancements (decolua/9router #3803 - #3954)

package/app/node_modules/@aws/lambda-invoke-store/dist-cjs/invoke-store.js

@@ -0,0 +1,131 @@
+'use strict';
+
+const PROTECTED_KEYS = {
+    REQUEST_ID: Symbol.for("_AWS_LAMBDA_REQUEST_ID"),
+    X_RAY_TRACE_ID: Symbol.for("_AWS_LAMBDA_X_RAY_TRACE_ID"),
+    TENANT_ID: Symbol.for("_AWS_LAMBDA_TENANT_ID"),
+    TRACEPARENT: Symbol.for("_AWS_LAMBDA_TRACEPARENT"),
+    TRACESTATE: Symbol.for("_AWS_LAMBDA_TRACESTATE"),
+    BAGGAGE: Symbol.for("_AWS_LAMBDA_BAGGAGE"),
+};
+const NO_GLOBAL_AWS_LAMBDA = ["true", "1"].includes(process.env?.AWS_LAMBDA_NODEJS_NO_GLOBAL_AWSLAMBDA ?? "");
+if (!NO_GLOBAL_AWS_LAMBDA) {
+    globalThis.awslambda = globalThis.awslambda || {};
+}
+class InvokeStoreBase {
+    static PROTECTED_KEYS = PROTECTED_KEYS;
+    isProtectedKey(key) {
+        return Object.values(PROTECTED_KEYS).includes(key);
+    }
+    getRequestId() {
+        return this.get(PROTECTED_KEYS.REQUEST_ID) ?? "-";
+    }
+    getXRayTraceId() {
+        return this.get(PROTECTED_KEYS.X_RAY_TRACE_ID);
+    }
+    getTenantId() {
+        return this.get(PROTECTED_KEYS.TENANT_ID);
+    }
+    getTraceparent() {
+        return this.get(PROTECTED_KEYS.TRACEPARENT);
+    }
+    getTracestate() {
+        return this.get(PROTECTED_KEYS.TRACESTATE);
+    }
+    getBaggage() {
+        return this.get(PROTECTED_KEYS.BAGGAGE);
+    }
+}
+class InvokeStoreSingle extends InvokeStoreBase {
+    currentContext;
+    getContext() {
+        return this.currentContext;
+    }
+    hasContext() {
+        return this.currentContext !== undefined;
+    }
+    get(key) {
+        return this.currentContext?.[key];
+    }
+    set(key, value) {
+        if (this.isProtectedKey(key)) {
+            throw new Error(`Cannot modify protected Lambda context field: ${String(key)}`);
+        }
+        this.currentContext = this.currentContext || {};
+        this.currentContext[key] = value;
+    }
+    run(context, fn) {
+        this.currentContext = context;
+        return fn();
+    }
+}
+class InvokeStoreMulti extends InvokeStoreBase {
+    als;
+    static async create() {
+        const instance = new InvokeStoreMulti();
+        const asyncHooks = await import('node:async_hooks');
+        instance.als = new asyncHooks.AsyncLocalStorage();
+        return instance;
+    }
+    getContext() {
+        return this.als.getStore();
+    }
+    hasContext() {
+        return this.als.getStore() !== undefined;
+    }
+    get(key) {
+        return this.als.getStore()?.[key];
+    }
+    set(key, value) {
+        if (this.isProtectedKey(key)) {
+            throw new Error(`Cannot modify protected Lambda context field: ${String(key)}`);
+        }
+        const store = this.als.getStore();
+        if (!store) {
+            throw new Error("No context available");
+        }
+        store[key] = value;
+    }
+    run(context, fn) {
+        return this.als.run(context, fn);
+    }
+}
+exports.InvokeStore = void 0;
+(function (InvokeStore) {
+    let instance = null;
+    async function getInstanceAsync(forceInvokeStoreMulti) {
+        if (!instance) {
+            instance = (async () => {
+                const isMulti = forceInvokeStoreMulti === true || "AWS_LAMBDA_MAX_CONCURRENCY" in process.env;
+                const newInstance = isMulti
+                    ? await InvokeStoreMulti.create()
+                    : new InvokeStoreSingle();
+                if (!NO_GLOBAL_AWS_LAMBDA && globalThis.awslambda?.InvokeStore) {
+                    return globalThis.awslambda.InvokeStore;
+                }
+                else if (!NO_GLOBAL_AWS_LAMBDA && globalThis.awslambda) {
+                    globalThis.awslambda.InvokeStore = newInstance;
+                    return newInstance;
+                }
+                else {
+                    return newInstance;
+                }
+            })();
+        }
+        return instance;
+    }
+    InvokeStore.getInstanceAsync = getInstanceAsync;
+    InvokeStore._testing = process.env.AWS_LAMBDA_BENCHMARK_MODE === "1"
+        ? {
+            reset: () => {
+                instance = null;
+                if (globalThis.awslambda?.InvokeStore) {
+                    delete globalThis.awslambda.InvokeStore;
+                }
+                globalThis.awslambda = { InvokeStore: undefined };
+            },
+        }
+        : undefined;
+})(exports.InvokeStore || (exports.InvokeStore = {}));
+
+exports.InvokeStoreBase = InvokeStoreBase;

package/app/node_modules/@aws/lambda-invoke-store/package.json

@@ -0,0 +1,55 @@
+{
+  "name": "@aws/lambda-invoke-store",
+  "version": "0.3.0",
+  "description": "Invoke scoped data storage for AWS Lambda Node.js Runtime Environment",
+  "homepage": "https://github.com/awslabs/aws-lambda-invoke-store",
+  "main": "./dist-cjs/invoke-store.js",
+  "types": "./dist-types/invoke-store.d.ts",
+  "module": "./dist-es/invoke-store.js",
+  "exports": {
+    ".": {
+      "types": "./dist-types/invoke-store.d.ts",
+      "module": "./dist-es/invoke-store.js",
+      "node": "./dist-cjs/invoke-store.js",
+      "import": "./dist-es/invoke-store.js",
+      "require": "./dist-cjs/invoke-store.js"
+    }
+  },
+  "files": [
+    "dist-es",
+    "dist-cjs",
+    "dist-types"
+  ],
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/awslabs/aws-lambda-invoke-store.git"
+  },
+  "license": "Apache-2.0",
+  "author": {
+    "name": "Amazon Web Services",
+    "url": "http://aws.amazon.com"
+  },
+  "scripts": {
+    "build": "yarn clean && yarn build:types && node ./scripts/build-rollup.js",
+    "build:types": "tsc -p tsconfig.types.json",
+    "clean": "rm -rf dist-types dist-cjs dist-es",
+    "test": "vitest run --reporter verbose",
+    "test:watch": "vitest watch",
+    "release": "yarn build && changeset publish"
+  },
+  "devDependencies": {
+    "@changesets/cli": "^2.29.6",
+    "@rollup/plugin-node-resolve": "^16.0.3",
+    "@rollup/plugin-typescript": "^12.3.0",
+    "@tsconfig/node18": "^18.2.4",
+    "@types/node": "^18.19.130",
+    "rollup": "^4.52.5",
+    "tslib": "^2.8.1",
+    "typescript": "^5.9.3",
+    "vitest": "^3.1.1"
+  },
+  "engines": {
+    "node": ">=18.0.0"
+  },
+  "packageManager": "yarn@4.9.4"
+}

package/app/node_modules/@aws-sdk/core/client.js

@@ -0,0 +1,5 @@
+/**
+ * Do not edit:
+ * This is a compatibility redirect for contexts that do not understand package.json exports field.
+ */
+module.exports = require("./dist-cjs/submodules/client/index.js");