@lcv-ideas-software/cross-review 4.0.0

package/docs/reports/cross-review-v2-token-streaming-smoke-2026-04-30.md

@@ -0,0 +1,119 @@
+# Cross Review v2 - Token Streaming Smoke
+
+Date: 2026-04-30, America/Sao_Paulo
+Runtime under test: local `cross-review-v2` source, package version `2.2.0`
+
+## Purpose
+
+This report records a real provider token-streaming smoke test for
+`cross-review-v2`. It verifies that the four peer adapters use provider-native
+streaming APIs and emit session token-progress events.
+
+No API key, prompt body or raw provider response text was printed or written to
+this report.
+
+## Command
+
+```powershell
+npm run api-streaming-smoke
+```
+
+## Result
+
+The command completed successfully.
+
+```json
+{
+  "ok": true,
+  "streaming_tokens": true,
+  "streaming_text": false,
+  "results": [
+    {
+      "peer": "codex",
+      "provider": "openai",
+      "requested_model": "gpt-5.5",
+      "reported_model": "gpt-5.5-2026-04-23",
+      "status": "NEEDS_EVIDENCE",
+      "decision_quality": "clean",
+      "delta_events": 126,
+      "streamed_chars": 681,
+      "usage": {
+        "input_tokens": 471,
+        "output_tokens": 653,
+        "total_tokens": 1124,
+        "reasoning_tokens": 516
+      }
+    },
+    {
+      "peer": "claude",
+      "provider": "anthropic",
+      "requested_model": "claude-opus-4-7",
+      "reported_model": "claude-opus-4-7",
+      "status": "READY",
+      "decision_quality": "clean",
+      "delta_events": 14,
+      "streamed_chars": 890,
+      "usage": {
+        "input_tokens": 1094,
+        "output_tokens": 308,
+        "total_tokens": 1402
+      }
+    },
+    {
+      "peer": "gemini",
+      "provider": "google",
+      "requested_model": "gemini-3.1-pro-preview",
+      "reported_model": "gemini-3.1-pro-preview",
+      "status": "READY",
+      "decision_quality": "clean",
+      "delta_events": 3,
+      "streamed_chars": 282,
+      "usage": {
+        "input_tokens": 391,
+        "output_tokens": 59,
+        "total_tokens": 1182,
+        "reasoning_tokens": 732
+      }
+    },
+    {
+      "peer": "deepseek",
+      "provider": "deepseek",
+      "requested_model": "deepseek-v4-pro",
+      "reported_model": "deepseek-v4-pro",
+      "status": "NEEDS_EVIDENCE",
+      "decision_quality": "clean",
+      "delta_events": 88,
+      "streamed_chars": 430,
+      "usage": {
+        "input_tokens": 402,
+        "output_tokens": 1561,
+        "total_tokens": 1963,
+        "reasoning_tokens": 1473
+      }
+    }
+  ]
+}
+```
+
+`NEEDS_EVIDENCE` in this smoke is not a failure. The prompt is intentionally
+minimal and artificial; the release criterion here is that each adapter streams
+provider output, emits token events, returns a parseable decision and reports
+usage/model metadata without leaking sensitive data.
+
+The default token-event mode records progress counts rather than raw provider
+text. `CROSS_REVIEW_V2_STREAM_TEXT=1` is available only as a trusted local
+diagnostic mode and still applies redaction before persistence.
+
+## Release Implications
+
+- `CROSS_REVIEW_V2_STREAM_TOKENS=1` is validated against real provider calls.
+- `CROSS_REVIEW_V2_STREAM_TOKENS=0` was validated through `npm run
+runtime-smoke`; `server_info.capabilities.token_streaming` and
+  `runtime_capabilities.capabilities.token_streaming` both reported `false`.
+- Token streaming is a real provider-output path, not simulated progress.
+- Existing final-result parsing remains active after streamed text is
+  accumulated.
+- Stub regression coverage verifies that summed streamed character counts match
+  the final parsed result text length.
+- The event stream can be consumed by MCP hosts and future UI layers to show
+  progress during slow reviews.

package/package.json

@@ -0,0 +1,88 @@
+{
+  "name": "@lcv-ideas-software/cross-review",
+  "version": "4.0.0",
+  "description": "API-first MCP server for multi-model cross-review with unanimous convergence gates.",
+  "type": "module",
+  "main": "dist/src/mcp/server.js",
+  "bin": {
+    "cross-review": "dist/src/mcp/server.js",
+    "cross-review-dashboard": "dist/src/dashboard/server.js"
+  },
+  "files": [
+    "dist/",
+    "docs/",
+    "README.md",
+    "LICENSE",
+    "NOTICE",
+    "SECURITY.md",
+    "CHANGELOG.md"
+  ],
+  "scripts": {
+    "build": "tsc -p tsconfig.json",
+    "clean": "node scripts/clean.mjs",
+    "prepack": "npm run clean && npm run build",
+    "dev": "tsx src/mcp/server.ts",
+    "dashboard": "tsx src/dashboard/server.ts",
+    "smoke": "tsx scripts/smoke.ts",
+    "runtime-smoke": "npm run build && tsx scripts/runtime-smoke.ts",
+    "runtime-default-smoke": "tsx scripts/runtime-default-smoke.ts",
+    "api-streaming-smoke": "tsx scripts/api-streaming-smoke.ts",
+    "test": "npm run build && npm run smoke && npm run runtime-smoke",
+    "lint": "eslint . --max-warnings=0",
+    "format": "prettier --write .",
+    "format:check": "prettier --check .",
+    "typecheck": "tsc -p tsconfig.json --noEmit",
+    "format:public": "prettier --write \"**/index.html\"",
+    "format:public:check": "prettier --check \"**/index.html\""
+  },
+  "dependencies": {
+    "@anthropic-ai/sdk": "^0.96.0",
+    "@google/genai": "^2.0.0",
+    "@modelcontextprotocol/sdk": "^1.29.0",
+    "openai": "^6.35.0",
+    "pino": "^10.3.1",
+    "zod": "^4.3.6"
+  },
+  "overrides": {
+    "ip-address": ">=10.1.1",
+    "hono": ">=4.12.16"
+  },
+  "devDependencies": {
+    "@eslint/js": "^10.0.1",
+    "@types/node": "^25.6.0",
+    "eslint": "^10.2.1",
+    "prettier": "^3.8.3",
+    "tsx": "^4.21.0",
+    "typescript": "^6.0.3",
+    "typescript-eslint": "^8.59.1"
+  },
+  "engines": {
+    "node": ">=22"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "license": "Apache-2.0",
+  "author": "LCV Ideas & Software",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/LCV-Ideas-Software/cross-review.git"
+  },
+  "bugs": {
+    "url": "https://github.com/LCV-Ideas-Software/cross-review/issues"
+  },
+  "homepage": "https://cross-review.lcv.dev/",
+  "keywords": [
+    "mcp",
+    "model-context-protocol",
+    "cross-review",
+    "multi-agent",
+    "openai",
+    "anthropic",
+    "gemini",
+    "deepseek",
+    "grok",
+    "xai",
+    "api"
+  ]
+}