tapd-bug-analyzer

TAPD Bug Analyzer Skill

Automatically analyze and locate issues in TAPD Bug tickets, supporting multiple analysis strategies.

Usage

/tapd-bug <bug_url> [--env <environment>] [--region <region>]

Parameters

bug_url: Required, TAPD Bug ticket link (e.g.:https://www.tapd.cn/xxxxx/bugtrace/bugs/view/xxxxxx）
--env: Optional, environment name: dev/test/prod (defaults to automatic determination based on URL)
--region: Optional, region: shanghai/hangzhou (default shanghai)

Instructions

When the user invokes this skill, execute the following steps:

Step 1: Parse the TAPD Bug link

Extract workspace_id and bug_id from the Bug link:

URL 格式: https://www.tapd.cn/{workspace_id}/bugtrace/bugs/view/{bug_id}
示例: https://www.tapd.cn/12345678/bugtrace/bugs/view/1012345678001234567

Step 2: Get Bug details

Use the TAPD MCP tool to get Bug ticket details:

调用 mcp__mcp-server-tapd__get_bug 工具：
- workspace_id: 从 URL 解析得到
- options: {"id": "<bug_id>", "fields": "id,title,description,status,severity,priority,reporter,current_owner,created,modified"}

Step 3: Analyze the Bug description and choose an analysis strategy

Carefully read the Bug description (description field), and choose an analysis strategy by priority:

Strategy A: Contains traceId

Check whether the Bug description contains a traceId (usually a 16-character hexadecimal string):

Regex match: [a-f0-9]{16}
Common format: traceId: xxx, trace_id=xxx, X-B3-TraceId: xxx

If a traceId is found, directly call sls-trace-analyzer:

~/.claude/skills/sls-trace-analyzer/.venv/bin/python \
  ~/.claude/skills/sls-trace-analyzer/sls-query.py \
  --trace-id "<traceId>" \
  --region "<region>" \
  --env "<env>"

Note: By default, the script queries both the regular storage and long-term storage logstores at the same time to obtain complete logs.

Strategy B: Contains request URL (with timestamp)

Check whether the Bug description contains an API request URL:

URL pattern: http(s)://xxx.abczs.cn/api/...
Timestamp parameter: the 13-digit number at the end of the URL (millisecond timestamp)

Example URL:

http://32.ftest.abczs.cn/api/v2/patientorders/his/ffffffff0000000035246dd0adec0000/hospital?1768198926187

If a URL with a timestamp is found, directly call sls-trace-analyzer (URL mode):

~/.claude/skills/sls-trace-analyzer/.venv/bin/python \
  ~/.claude/skills/sls-trace-analyzer/sls-query.py \
  --url "<完整URL含时间戳>"

sls-trace-analyzer will automatically:

Parse the domain, path, and timestamp in the URL
Determine the environment (dev/test/prod) and region based on the domain
Query gateway logs to obtain the traceId
Use the traceId to query the complete call chain logs
Output the log analysis results

Strategy C: Neither traceId nor timestamp

Perform code analysis based on the Bug description:

Extract key information (error messages, functional modules, operation steps)
Use codebase-retrieval to search for related code
Analyze the business logic and infer the cause of the issue

Step 4: Log analysis and issue localization

After obtaining the traceId, call sls-trace-analyzer for in-depth analysis:

Identify errors and exceptions
- Find ERROR and WARN level logs
- Identify Java exception stack traces (Exception, Caused by)
- Extract error messages and error codes
Trace the request chain
- Sort log entries by time
- Identify the request entry point (Controller) and exit point
- Mark RPC/Feign calls and responses
Locate code issues
- Extract class names and line numbers from exception stack traces
- Prioritize concern with cn.abcyun code under the package
- Use codebase-retrieval to find the corresponding code

Step 5: Output the analysis report

Generate a structured analysis report:

## TAPD Bug 分析报告

### Bug 基本信息
- **Bug ID**: xxx
- **标题**: xxx
- **状态**: xxx
- **严重程度**: xxx
- **报告人**: xxx
- **处理人**: xxx

### 分析策略
- 使用策略: A/B/C
- TraceId: xxx（如有）
- 请求时间: xxx（如有）

### 错误摘要
- **错误类型**: xxx
- **错误消息**: xxx
- **发生位置**: `文件路径:行号`

### 调用链路
1. [入口] POST /api/xxx
2. [Service] XxxService.method()
3. [RPC] 调用 xxx 服务
4. [Error] 发生异常

### 问题定位
- **根因分析**: xxx
- **相关代码**:
  ```java
  // 问题代码片段

Fix suggestions


## Examples

### 示例 1: Bug 描述中包含 traceId

```bash
/tapd-bug https://www.tapd.cn/12345678/bugtrace/bugs/view/1012345678001234567

Bug description content:

用户反馈下单失败，错误信息：系统繁忙，请稍后重试
traceId: 8989a698c36e5189

Execution flow:

Parse the URL to obtain workspace_id=12345678, bug_id=1012345678001234567
Call TAPD MCP to get Bug details
Extract traceId from the description: 8989a698c36e5189
Call sls-trace-analyzer to analyze logs
Output analysis report

Example 2: The Bug description contains a request URL

/tapd-bug https://www.tapd.cn/12345678/bugtrace/bugs/view/1012345678001234568 --env test

Bug description content:

测试环境下单接口报错 500
请求地址：http://32.ftest.abczs.cn/api/v2/patientorders/his/ffffffff0000000035246dd0adec0000/hospital?1768198926187

Execution flow:

Parse the URL to get workspace_id and bug_id
Call TAPD MCP to get Bug details
Extract the request URL (including timestamp) from the description
Call sls-trace-analyzer (URL mode) for log analysis
Output analysis report

Example 3: A Bug described in plain text

/tapd-bug https://www.tapd.cn/12345678/bugtrace/bugs/view/1012345678001234569

Bug description content:

患者在门诊缴费时，选择医保支付后点击确认，页面提示"医保结算失败"，
但实际上 HIS 系统显示已经结算成功。
复现步骤：
1. 登录患者端 APP
2. 选择待缴费订单
3. 选择医保支付
4. 点击确认支付
预期结果：支付成功，跳转到支付成功页面
实际结果：提示"医保结算失败"

Execution flow:

Parse the URL to get workspace_id and bug_id
Call TAPD MCP to get Bug details
No traceId or timestamp found
Analyze the description to extract keywords: outpatient payment, medical insurance payment, medical insurance settlement
Use codebase-retrieval to search related code
Analyze the business logic and infer possible causes
Output analysis report

Example 4: Real case - prescription does not display specification

Input:

/tapd-bug https://www.tapd.cn/43780818/bugtrace/bugs/view/1143780818001109821

Bug description content:

问题概述：处方不展示规格
门店id: ffffffff000000003526121e726a0004
收费单id：ffffffff000000003526dc97b673c003
补充说明：后台没返回药品档案的规格

Execution flow:

Parse URL: workspace_id=43780818, bug_id=1143780818001109821
Call TAPD MCP to get Bug details
Analyze the description: no traceId, no timestamp URL → use strategy C
Extract keywords: prescription, specification, drug profile
Use codebase-retrieval to search related code
Located at PrescriptionService.java:678-680

Problem localization result:

// PrescriptionService.java:678-680
if (goodsItem.getType() == 2) {
    prescriptionFormItem.setSpecification(...);
}

Root cause analysis: The specification field of Western medicine prescriptions is only set when type==2 (materials); the specification is not set for ordinary Western medicines (type=1).

Fix suggestion: Set the specification for all drug types, preferring the GoodsItem.displaySpec field.

About this skill

Usage

Parameters

Instructions

Step 1: Parse the TAPD Bug link

Step 2: Get Bug details

Step 3: Analyze the Bug description and choose an analysis strategy

Strategy A: Contains traceId

Strategy B: Contains request URL (with timestamp)

Strategy C: Neither traceId nor timestamp

Step 4: Log analysis and issue localization

Step 5: Output the analysis report

Fix suggestions

Example 2: The Bug description contains a request URL

Example 3: A Bug described in plain text

Example 4: Real case - prescription does not display specification