feat(core): unify error codes and messages on a bilingual error catalog

.github/workflows/ci.yml

@@ -30,109 +30,3 @@ jobs:
 
       - name: Build and run tests
         run: mvn -B clean verify
-
-  e2e-test:
-    name: E2E Samples Test
-    runs-on: ubuntu-latest
-    needs: build-and-test
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-
-      - name: Set up JDK 17
-        uses: actions/setup-java@v4
-        with:
-          distribution: temurin
-          java-version: '17'
-          cache: maven
-
-      - name: Package project
-        run: mvn -B package -DskipTests
-
-      - name: Create server env
-        run: |
-          cat > server.env << 'EOF'
-          A2AT_LANGUAGE=zh-CN
-          A2AT_PROMPT_SOURCE_TYPE=local_file
-          A2AT_PROMPT_RESOURCE_LOCAL_ROOT_DIR=a2a-t-resources/src/main/resources/prompt_resources
-          A2AT_LLM_PROVIDER=openai
-          A2AT_LLM_MODEL=deepseek-chat
-          A2AT_LLM_BASE_URL=https://api.deepseek.com
-          A2AT_LLM_API_KEY=
-          A2AT_NEGOTIATION_STATE_STORE_TYPE=in_memory
-          A2AT_SAMPLE_HOST=127.0.0.1
-          A2AT_SAMPLE_PORT=26335
-          REGISTRY_CENTER_HOST=127.0.0.1
-          REGISTRY_CENTER_PORT=5001
-          EOF
-          sed -i 's|^A2AT_LLM_API_KEY=$|A2AT_LLM_API_KEY=${{ secrets.A2AT_LLM_API_KEY }}|' server.env
-
-      - name: Create client env
-        run: |
-          cat > client.env << 'EOF'
-          A2AT_LANGUAGE=zh-CN
-          A2AT_PROMPT_SOURCE_TYPE=classpath
-          A2AT_PROMPT_RESOURCE_LOCAL_ROOT_DIR=
-          A2AT_LLM_PROVIDER=openai
-          A2AT_LLM_MODEL=deepseek-chat
-          A2AT_LLM_BASE_URL=https://api.deepseek.com
-          A2AT_LLM_API_KEY=
-          A2AT_NEGOTIATION_STATE_STORE_TYPE=in_memory
-          A2AT_SAMPLE_HOST=127.0.0.1
-          A2AT_SAMPLE_PORT=26335
-          EOF
-          sed -i 's|^A2AT_LLM_API_KEY=$|A2AT_LLM_API_KEY=${{ secrets.A2AT_LLM_API_KEY }}|' client.env
-
-      - name: Diagnose LLM config availability
-        shell: bash
-        run: |
-          line=$(grep 'A2AT_LLM_API_KEY=' server.env)
-          value=${line#A2AT_LLM_API_KEY=}
-          len=${#value}
-          if [ "$len" -eq 0 ]; then
-            echo "::notice::A2AT_LLM_API_KEY is empty, e2e will use mock LLM responses"
-          else
-            prefix=$(echo "$value" | cut -c1-12)
-            echo "A2AT_LLM_API_KEY length=$len prefix=$prefix..."
-          fi
-
-      - name: Start server
-        run: |
-          nohup java @a2a-t-sample/target/server.javaargs.txt > server.log 2>&1 &
-          echo $! > server.pid
-          sleep 3
-
-      - name: Wait for server
-        run: |
-          for i in $(seq 1 10); do
-            if curl -s http://127.0.0.1:26335/ > /dev/null 2>&1; then
-              echo "Server ready"
-              break
-            fi
-            echo "Waiting for server... attempt $i"
-            sleep 2
-          done
-
-      - name: Run client
-        env:
-          A2AT_SAMPLE_MAX_ARTIFACTS: "5"
-        run: |
-          java @a2a-t-sample/target/client.javaargs.txt 2>&1 | tee client.log
-          artifact_count=$(grep -c 'task-artifact' client.log || true)
-          echo "e2e artifacts received: $artifact_count"
-          if [ "$artifact_count" -eq 0 ]; then
-            echo "::error::e2e failed: no artifacts received"
-            exit 1
-          fi
-          echo "e2e PASSED"
-
-      - name: Show logs
-        if: always()
-        run: |
-          echo "=== server.log ===" && cat server.log 2>/dev/null || true
-          echo "=== client.log ===" && cat client.log 2>/dev/null || true
-
-      - name: Cleanup
-        if: always()
-        run: |
-          kill $(cat server.pid 2>/dev/null) 2>/dev/null || true

.gitignore

@@ -3,9 +3,19 @@ target/
 .gitignore
 .flattened-pom.xml
 *.env
+# Sample env templates carry no secrets (empty API key) and must ship with the sources.
+!a2a-t-sample/src/main/resources/**/negotiation.env
+!a2a-t-sample/src/main/resources/**/authz.env
 reference/
 findings.md
 progress.md
 task_plan.md
 
 .idea/
+
+.worktrees/
+eval-results/
+eval-run*.log
+eval-smoke.json
+.jqwik-database
+__pycache__/

README.md

@@ -51,7 +51,7 @@ This SDK is primarily aimed at two types of users:
 - Client API: Provides a task prompt generation result stream, along with negotiation entry points such as `start_negotiation`, `receive_negotiation`, and `continue_negotiation`.
 - Server validation API: Targets A2A-T protocol messages that conform to the SDK format, performing metadata parsing, slot extraction.
 - Negotiation types: Includes four built-in negotiation types: `information`.
-- Negotiation-T content layer: Template-driven negotiation message generation (`fromData` deterministic, `fromText` with one LLM extraction step), compliance checking with parameter extraction (`validateAndFilling*Data`), and negotiation template queries. See the [Developer Guide §1.10](docs/en/developer_guide.md#110-negotiation-t-content-layer).
+- Negotiation-T content layer: Template-driven negotiation message generation (`fromData` deterministic, `fromText` with one LLM extraction step), compliance checking with parameter extraction (`validatePromptAndDataFilling*Data`), and negotiation template queries. See the [Developer Guide §1.10](docs/en/developer_guide.md#110-negotiation-t-content-layer).
 - Resource organization: Built-in prompt resources are located in `a2a-t-resources/src/main/resources/prompt_resources`, containing `prompts`, `scenarios`, `slots`, and `templates`.
 - Built-in example scenario: Currently, the package provides scenarios such as`subscribe_incident`.
 
@@ -67,6 +67,7 @@ The repository is organized as a Maven multi-module reactor, with core code unde
 - `a2a-t-negotiation`: Negotiation types, runtime state machine, and state storage.
 - `a2a-t-client`: Client facade providing task prompt generation and negotiation entry points.
 - `a2a-t-server`: Server facade providing `processed task prompt` validation and negotiation entry points.
+- `a2a-t-corpus`: A2A-T conformance test corpus module — a pure test module with data-driven corpus cases and property-based tests.
 - `a2a-t-sample`: Runnable client/server sample.
 
 ## Installation and Environment Requirements

README_zh.md

@@ -51,7 +51,7 @@ SPDX-License-Identifier: Apache-2.0
 - 客户端 API：提供任务提示词生成结果流，以及 `start_negotiation`、`receive_negotiation`、`continue_negotiation` 等协商入口。
 - 服务端校验 API：面向符合 SDK 格式的 `processed task prompt`，执行元数据解析、槽位提取和槽位校验。
 - 协商类型：内置 `information` 协商类型。
-- Negotiation-T 内容层：以模板为驱动的协商消息生成（`fromData` 确定性执行、`fromText` 含一次 LLM 抽取）、合规校验与参数提取（`validateAndFilling*Data`）以及协商模板查询。参见[开发指南 §1.10](docs/zh/开发指南.md)。
+- Negotiation-T 内容层：以模板为驱动的协商消息生成（`fromData` 确定性执行、`fromText` 含一次 LLM 抽取）、合规校验与参数提取（`validatePromptAndDataFilling*Data`）以及协商模板查询。参见[开发指南 §1.10](docs/zh/开发指南.md)。
 - 资源组织：内置提示词资源位于 `a2a-t-resources/src/main/resources/prompt_resources`，包含 `prompts`、`scenarios`、`slots`、`templates`。
 - 内置示例场景：当前随包提供 `subscribe_incident`等场景。
 
@@ -67,6 +67,7 @@ SPDX-License-Identifier: Apache-2.0
 - `a2a-t-negotiation`：协商类型、运行时状态机与状态存储。
 - `a2a-t-client`：客户端封装，提供任务提示词生成与协商入口。
 - `a2a-t-server`：服务端封装，提供A2A-T协议报文的校验与协商入口。
+- `a2a-t-corpus`：A2A-T 一致性测试语料模块——纯测试模块，含数据驱动用例与属性测试。
 - `a2a-t-sample`：可运行的客户端/服务端示例。
 
 ## 安装与环境要求