Add Jev Browser (openqa-cn) listing

README.md

@@ -289,6 +289,7 @@ The [independent model research](community/projects/tools/README.md#independent-
 - [AnchorLint](https://github.com/prantikmedhi/anchorlint) - Audits internal links in built HTML sites with deterministic checks plus optional TypeSafe Jev promise/relevance judgments. [Project guide](community/projects/tools/anchorlint.md).
 - [jev-seo](https://github.com/AkashPriyadarshii/jev-seo) - Agent-oriented SEO/GEO CLI and MCP with local DuckDuckGo helpers and optional TypeSafe Jev scoring. [Project guide](community/projects/tools/jev-seo.md).
 - [jev-seo (AgriciDaniel)](https://github.com/AgriciDaniel/jev-seo) - Live website SEO audit from one URL with TypeSafe Jev judgments and PDF/XLSX/Markdown reports (distinct from AkashPriyadarshii/jev-seo). [Project guide](community/projects/tools/agrici-jev-seo.md).
+- [Jev Browser (openqa-cn)](https://github.com/openqa-cn/jev-browser) - Indexed Playwright automation: TypeSafe Jev chooses control/op; CodexQA skill with replay and HTML reports. [Project guide](community/projects/tools/openqa-jev-browser.md).
 - [Jev Browser (tontoko)](https://github.com/tontoko/jev-browser) - Integrates Jev field selection and source-backed extraction with a Playwright SDK, CLI, MCP server, and explicit assertions. [Project guide](community/projects/tools/jev-browser-tontoko.md).
 - [Jev Browser (Ying-Kai-Liao)](https://github.com/Ying-Kai-Liao/jev-browser) - Turns caller-defined browser goals into Jev-selected Playwright actions with a library, CLI, and MCP server. [Project guide](community/projects/tools/jev-browser-ying-kai-liao.md).
 - [Jev Browser Skill](https://github.com/zurfyx/jev-browser-skill) - Claude Code/Codex skill where TypeSafe Jev chooses browser operations and targets from a code-built element table (reference implementation; explainer site included). [Project guide](community/projects/tools/jev-browser-skill.md).

community/projects/tools/README.md

@@ -25,6 +25,7 @@ See the [computer-use guide](../../../docs/computer-use.md) for a comparison, fo
 | [Cua jev-use](cua-jev-use.md) | Compose a bounded chooser with Driver and independent fixture verification. | Python / TypeScript · integration recipe |
 | [Footwork](footwork.md) | Dual-process browser agent: TypeSafe Jev as System 1 in front of browser-use System 2, with a code-owned arbiter and evidence verification. | Python/Rust · package (`jevdual` 0.0.1) |
 | [gpui-agent](gpui-agent.md) | Drive instrumented GPUI apps via accessibility: TypeSafe Jev chooses typed actions/targets (no screenshots to the model). | Rust/Python/TypeScript · experimental native toolkit |
+| [Jev Browser (openqa-cn)](openqa-jev-browser.md) | Indexed Playwright automation: TypeSafe Jev chooses control/op; replay, generate, explore, HTML reports (CodexQA skill). | TypeScript · CLI (`codexqa-jev-browser` 0.1.0, MIT) |
 | [Jev Browser (tontoko)](jev-browser-tontoko.md) | Fill forms, extract records with evidence, and add semantic selection to Playwright tests. | TypeScript · SDK, CLI and MCP |
 | [Jev Browser (Ying-Kai-Liao)](jev-browser-ying-kai-liao.md) | Run small browser goals with Jev action/target selection and direct inspection; page data and supplied values reach TypeSafe. | JavaScript · Playwright library, CLI and MCP |
 | [Jev Browser Skill](jev-browser-skill.md) | Learn a minimal Jev-driven browser loop as a Claude Code/Codex skill (reference; see Ultrafast for fuller agents). | Agent skill + CDP scripts (`scripts/*.mjs`); explainer site |

community/projects/tools/openqa-jev-browser.md

@@ -0,0 +1,49 @@
+# Jev Browser (openqa-cn)
+
+[All projects](../README.md) · [Browser and computer use](README.md#browser-and-computer-use)
+
+Indexed Playwright browser automation: the page builds a closed control index; TypeSafe Jev chooses operation and target; Playwright acts. CodexQA skill with replay, generate, explore, and HTML reports.
+
+| At a glance | Details |
+| --- | --- |
+| Source | [Source](https://github.com/openqa-cn/jev-browser) |
+| Maintainer | [openqa-cn](https://github.com/openqa-cn). Independently curated; this entry is not an upstream submission or endorsement. |
+| Format | Node.js CLI **`codexqa-jev-browser` 0.1.0** + `SKILL.md` (also listed in CodexQA catalog). |
+| Requirements | Node.js ≥ 20; Playwright Chromium; `TYPESAFE_API_KEY` for live auto/generate (OpenAI-compatible chat is fallback planner/decision). Offline `observe` / `run` / `--decisions` need no model. |
+| License | [MIT](https://github.com/openqa-cn/jev-browser/blob/ea82bc96597b07f27a386288b196b0bca1cdcb3a/LICENSE). |
+| Disclosure | AI-assisted catalog review; no affiliation. Listing is not an endorsement. Source inspected (README, LICENSE, package). Live browser/Jev runs were **not** executed on the review host. TypeSafe latency/cost comparisons in the upstream README are vendor-reported, not reproduced here. |
+
+## When to use
+
+Use it when you want index-grounded browser steps (no screenshot-to-coordinates planner) with YAML replay after `generate`. Prefer [Jev Ultrafast](jev-ultrafast.md) / [Midscene JEV Runner](midscene-jev-runner.md) for other Playwright+Jev stacks; prefer [Jev Browser Skill](jev-browser-skill.md) for a minimal teaching skill.
+
+## How it works
+
+Observation assigns each visible control an id, role, name, and allowed ops. Live steps ask Jev `/systemone` which op and which index (typing phrases come from the goal). Actions hit `data-codexqa-jev-browser-id`. Knowledge notes bias decisions; `--decisions` scripts skip the model. Reports keep marked screenshots and timing under `reports/`.
+
+## Get started
+
+```sh
+git clone https://github.com/openqa-cn/jev-browser.git
+cd jev-browser
+git checkout ea82bc96597b07f27a386288b196b0bca1cdcb3a
+npm install
+npx codexqa-jev-browser observe examples/app/index.html
+npx codexqa-jev-browser run cases/examples/search-docs.yaml
+# Live auto needs .env with TYPESAFE_API_KEY (charges)
+```
+
+## Examples and demos
+
+- `cases/examples/`, `cases/scripts/decisions-search.yaml`, `examples/app/` local HTML.
+- HTML report preview linked from README; offline `tests/` (vitest) do not call live models.
+
+## Limits and data handling
+
+Live decisions send page index text and goal phrases to TypeSafe or the chat fallback; password field values are omitted from model requests. Knowledge notes are required for site-specific facts. This listing did not run headed Chromium or live Jev.
+
+## Review and maintenance
+
+Reviewed on **2026-09-23** at [commit ea82bc9](https://github.com/openqa-cn/jev-browser/tree/ea82bc96597b07f27a386288b196b0bca1cdcb3a) (**0.1.0**, MIT). AI-assisted source review of README, LICENSE, package metadata. No live TypeSafe spend.
+
+Related: [Midscene JEV Runner](midscene-jev-runner.md), [Jev Browser Skill](jev-browser-skill.md), [Jev Ultrafast](jev-ultrafast.md).