Update dataset card changelog

README.md

@@ -24,6 +24,7 @@ size_categories:
   <a href="#leaderboard">LeaderBoard</a> •
   <a href="#distribution">Distribution</a> •
   <a href="#quick-start">Quick Start</a> •
+  <a href="#changelog">Changelog</a> •
   <a href="#citation">Citation</a>
 </div>
 
@@ -78,6 +79,19 @@ The figure above shows a leaderboard snapshot of rubric pass rates on Workspace-
 
 The distribution figure provides a compact overview of the Lite split, including file types, task abilities, task difficulty, workspace allocation, rubric counts, required files per task, and dependency structure. Although lighter than the full benchmark, Workspace-Bench-Lite is designed to preserve the key distributional characteristics of Workspace-Bench.
 
+## Repository Contents
+
+This repository provides the cleaned task metadata release for Workspace-Bench-Lite. It is paired with the workspace filesystem archives in [Workspace-Bench-Workspaces](https://huggingface.co/datasets/Workspace-Bench/Workspace-Bench-Workspaces), rather than storing the large raw workspace contents in this repository.
+
+Current contents include:
+
+- `task_lite_clean_en/`: cleaned English Lite task metadata organized by `absolute_id`
+- `task_lite_clean_cn/`: cleaned Chinese Lite task metadata organized by `absolute_id`
+- `task_lite_clean_en_metadata_table.csv`: a flat English Lite metadata table for analysis and filtering
+- `task_lite_clean_cn_metadata_table.csv`: a flat Chinese Lite metadata table for analysis and filtering
+
+At the time of writing, each language split contains **100 tasks**, one folder per task ID. Task metadata files include a `language` field: `"en"` for English tasks and `"cn"` for Chinese tasks.
+
 ## Quick Start
 
 Please refer to the official GitHub repository for benchmark setup, evaluation instructions, and updates:
@@ -86,6 +100,15 @@ Please refer to the official GitHub repository for benchmark setup, evaluation i
 
 More Lite-specific usage notes will be added in future updates.
 
+## Changelog
+
+### 2026-07-02
+
+- Added a `language` field to task `metadata.json` files and metadata table CSV files. The value is `"en"` for English tasks and `"cn"` for Chinese tasks.
+- Tasks `95`, `328`, and `388`: placed the required input files explicitly under their designated paths in the workspace filesystem, so agents can find the files from the workspace context.
+- Tasks `288`, `289`, and `258`: replaced corrupted input files with readable workspace files.
+- Tasks `289`, `274`, `334`, and `108`: refined selected rubrics to make the evaluation criteria more precise, self-consistent, and aligned with the expected output format.
+
 ## Citation
 
 If you use Workspace-Bench-Lite in your research, please cite the Workspace-Bench paper: