Process raw feedback into actionable, well-documented GitHub issues. Every feedback item is investigated before issue creation.
Core principle: Investigate first, issue second. Never create an issue without understanding what you're documenting.
Announce at start: "I'm using feedback-triage to investigate and create issues from this feedback."
将原始反馈转化为可执行、文档完善的GitHub问题。在创建问题之前,会对每一条反馈事项进行调查。
核心原则:先调查,再创建问题。在不了解要记录的内容之前,绝不创建问题。
开始时告知:“我正在使用feedback-triage来调查这些反馈并创建对应的问题。”
When to Use This Skill
何时使用该技能
Use this skill when you receive:
| Trigger | Examples |
|---|
| UAT feedback | "We have bugs from UAT testing..." |
| User testing results | "Users reported the following issues..." |
| Bug reports | "Here are the errors we found..." |
| Stakeholder feedback | "The client wants these changes..." |
| QA findings | "QA discovered these problems..." |
| Support escalations | "Support tickets about..." |
| Production incidents | "These errors are occurring in prod..." |
| Feature requests batch | "Users have requested..." |
| UX review findings | "The UX review identified..." |
Key indicators:
- Multiple items in one message
- Raw feedback that needs investigation
- Error logs, curl commands, or screenshots
- Requests to "create issues" from feedback
- Phrases like "bugs to resolve", "issues from UAT", "feedback to triage"
当你收到以下内容时,使用本技能:
| 触发场景 | 示例 |
|---|
| UAT反馈 | "我们有来自UAT测试的bug..." |
| 用户测试结果 | "用户反馈了以下问题..." |
| Bug报告 | "这是我们发现的错误..." |
| 利益相关者反馈 | "客户希望做出这些更改..." |
| QA发现 | "QA发现了这些问题..." |
| 支持工单升级 | "关于...的支持工单" |
| 生产环境事故 | "生产环境中出现了这些错误..." |
| 批量功能请求 | "用户请求了..." |
| UX评审发现 | "UX评审指出了..." |
关键识别指标:
- 一条消息中包含多个事项
- 需要调查的原始反馈
- 错误日志、curl命令或截图
- 要求“根据反馈创建问题”
- 出现“需要修复的bug”、“来自UAT的问题”、“待分类的反馈”等表述
The Triage Protocol
分类处理流程
┌─────────────────────────────────────────────────────────────────────┐
│ FEEDBACK RECEIVED │
└─────────────────────────────────┬───────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────────┐
│ STEP 0: PROJECT BOARD READINESS (GATE) │
│ Verify GITHUB_PROJECT_NUM and GH_PROJECT_OWNER are set │
└─────────────────────────────────┬───────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────────┐
│ STEP 1: PARSE INTO ITEMS │
│ Identify distinct feedback items │
│ Create TodoWrite entry for each item │
└─────────────────────────────────┬───────────────────────────────────┘
│
▼
┌─────────────────────────────────────────┐
│ FOR EACH FEEDBACK ITEM │
└─────────────────────┬───────────────────┘
│
┌─────────────────────────┼─────────────────────────┐
▼ ▼ ▼
┌───────────┐ ┌─────────────┐ ┌─────────────┐
│INVESTIGATE│ │ CLASSIFY │ │ CREATE │
│ │──────────▶│ │──────────▶│ ISSUE │
│ Research │ │ Type+Priority│ │ Best Practice│
└───────────┘ └─────────────┘ └──────┬──────┘
│
▼
┌─────────────┐
│ ADD TO │
│ PROJECT │
│ BOARD │
└─────────────┘
┌─────────────────────────────────────────────────────────────────────┐
│ 收到反馈 │
└─────────────────────────────────┬───────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────────┐
│ 步骤0:项目看板就绪检查(准入门槛) │
│ 验证GITHUB_PROJECT_NUM和GH_PROJECT_OWNER已配置 │
└─────────────────────────────────┬───────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────────┐
│ 步骤1:拆分反馈为独立事项 │
│ 识别每个独立的反馈事项 │
│ 为每个事项创建TodoWrite条目 │
└─────────────────────────────────┬───────────────────────────────────┘
│
▼
┌─────────────────────────────────────────┐
│ 针对每个反馈事项执行 │
└─────────────────────┬───────────────────┘
│
┌─────────────────────────┼─────────────────────────┐
▼ ▼ ▼
┌───────────┐ ┌─────────────┐ ┌─────────────┐
│调查分析│ │ 分类 │ │ 创建问题 │
│ │──────────▶│ │──────────▶│ 最佳实践│
│ 调研 │ │ 类型+优先级│ │ │
└───────────┘ └─────────────┘ └──────┬──────┘
│
▼
┌─────────────┐
│ 添加至 │
│ 项目 │
│ 看板 │
└─────────────┘
Step 0: Project Board Readiness (GATE)
步骤0:项目看板就绪检查(准入门槛)
Before any triage, verify project board infrastructure is ready.
在开始任何分类处理之前,先验证项目看板基础设施已准备就绪。
Derive defaults from GITHUB_PROJECT if provided
Derive defaults from GITHUB_PROJECT if provided
if [ -z "$GITHUB_PROJECT_NUM" ] && [ -n "$GITHUB_PROJECT" ]; then
NUM_CANDIDATE=$(echo "$GITHUB_PROJECT" | sed -E 's#./projects/([0-9]+).#\1#')
if [ -n "$NUM_CANDIDATE" ] && [ "$NUM_CANDIDATE" != "$GITHUB_PROJECT" ]; then
export GITHUB_PROJECT_NUM="$NUM_CANDIDATE"
echo "Derived GITHUB_PROJECT_NUM=$GITHUB_PROJECT_NUM from GITHUB_PROJECT"
fi
fi
if [ -z "$GH_PROJECT_OWNER" ] && [ -n "$GITHUB_OWNER" ]; then
export GH_PROJECT_OWNER="$GITHUB_OWNER"
echo "Derived GH_PROJECT_OWNER=$GH_PROJECT_OWNER from GITHUB_OWNER"
fi
if [ -z "$GH_PROJECT_OWNER" ] && [ -n "$GITHUB_PROJECT" ]; then
OWNER_CANDIDATE=$(echo "$GITHUB_PROJECT" | sed -E 's#
https://github.com/(orgs|users)/([^/]+)/projects/[0-9]+#\2#')
if [ -n "$OWNER_CANDIDATE" ] && [ "$OWNER_CANDIDATE" != "$GITHUB_PROJECT" ]; then
export GH_PROJECT_OWNER="$OWNER_CANDIDATE"
echo "Derived GH_PROJECT_OWNER=$GH_PROJECT_OWNER from GITHUB_PROJECT"
fi
fi
if [ -z "$GH_PROJECT_OWNER" ]; then
REMOTE_URL=$(git remote get-url origin 2>/dev/null || true)
OWNER_CANDIDATE=$(echo "$REMOTE_URL" | sed -E 's#(git@|https://)github.com[:/]+([^/]+)/[^/]+(.git)?#\2#')
if [ -n "$OWNER_CANDIDATE" ] && [ "$OWNER_CANDIDATE" != "$REMOTE_URL" ]; then
export GH_PROJECT_OWNER="$OWNER_CANDIDATE"
echo "Derived GH_PROJECT_OWNER=$GH_PROJECT_OWNER from git remote"
fi
fi
if [ -z "$GITHUB_PROJECT_NUM" ] && [ -n "$GITHUB_PROJECT" ]; then
NUM_CANDIDATE=$(echo "$GITHUB_PROJECT" | sed -E 's#./projects/([0-9]+).#\1#')
if [ -n "$NUM_CANDIDATE" ] && [ "$NUM_CANDIDATE" != "$GITHUB_PROJECT" ]; then
export GITHUB_PROJECT_NUM="$NUM_CANDIDATE"
echo "Derived GITHUB_PROJECT_NUM=$GITHUB_PROJECT_NUM from GITHUB_PROJECT"
fi
fi
if [ -z "$GH_PROJECT_OWNER" ] && [ -n "$GITHUB_OWNER" ]; then
export GH_PROJECT_OWNER="$GITHUB_OWNER"
echo "Derived GH_PROJECT_OWNER=$GH_PROJECT_OWNER from GITHUB_OWNER"
fi
if [ -z "$GH_PROJECT_OWNER" ] && [ -n "$GITHUB_PROJECT" ]; then
OWNER_CANDIDATE=$(echo "$GITHUB_PROJECT" | sed -E 's#
https://github.com/(orgs|users)/([^/]+)/projects/[0-9]+#\2#')
if [ -n "$OWNER_CANDIDATE" ] && [ "$OWNER_CANDIDATE" != "$GITHUB_PROJECT" ]; then
export GH_PROJECT_OWNER="$OWNER_CANDIDATE"
echo "Derived GH_PROJECT_OWNER=$GH_PROJECT_OWNER from GITHUB_PROJECT"
fi
fi
if [ -z "$GH_PROJECT_OWNER" ]; then
REMOTE_URL=$(git remote get-url origin 2>/dev/null || true)
OWNER_CANDIDATE=$(echo "$REMOTE_URL" | sed -E 's#(git@|https://)github.com[:/]+([^/]+)/[^/]+(.git)?#\2#')
if [ -n "$OWNER_CANDIDATE" ] && [ "$OWNER_CANDIDATE" != "$REMOTE_URL" ]; then
export GH_PROJECT_OWNER="$OWNER_CANDIDATE"
echo "Derived GH_PROJECT_OWNER=$GH_PROJECT_OWNER from git remote"
fi
fi
Verify environment variables
Verify environment variables
if [ -z "$GITHUB_PROJECT_NUM" ]; then
echo "BLOCKED: GITHUB_PROJECT_NUM not set"
exit 1
fi
if [ -z "$GH_PROJECT_OWNER" ]; then
echo "BLOCKED: GH_PROJECT_OWNER not set"
exit 1
fi
if [ -z "$GITHUB_PROJECT_NUM" ]; then
echo "BLOCKED: GITHUB_PROJECT_NUM not set"
exit 1
fi
if [ -z "$GH_PROJECT_OWNER" ]; then
echo "BLOCKED: GH_PROJECT_OWNER not set"
exit 1
fi
Verify project is accessible
Verify project is accessible
gh project view "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER" --format json > /dev/null 2>&1
**Skill:** `project-board-enforcement`
---
gh project view "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER" --format json > /dev/null 2>&1
**技能依赖:** `project-board-enforcement`
---
Step 1: Parse Feedback into Items
步骤1:将反馈拆分为独立事项
Identify Distinct Items
识别独立事项
Read through the feedback and identify each distinct item. Look for:
- Separate headings or sections
- Numbered lists
- Different error messages or behaviors
- Distinct feature requests or changes
通读反馈内容,识别每个独立的事项。注意寻找:
- 独立的标题或章节
- 编号列表
- 不同的错误信息或行为描述
- 不同的功能请求或更改需求
Create Tracking List
创建跟踪列表
Use TodoWrite to track each item
Use TodoWrite to track each item
Example: 3 items from UAT feedback
Example: 3 items from UAT feedback
Create a summary table for the user:
Feedback Items Identified
已识别的反馈事项
| # | Summary | Type (Preliminary) | Severity |
|---|
| 1 | Family page error | Bug | High |
| 2 | Terminology needs review | UX/Research | Medium |
| 3 | Cannot add care recipient | Bug | High |
I will investigate each item before creating issues.
| 序号 | 摘要 | 初步类型 | 严重程度 |
|---|
| 1 | 家庭页面错误 | Bug | 高 |
| 2 | 术语需要审核 | UX/调研 | 中 |
| 3 | 无法添加被照护人 | Bug | 高 |
我会在创建问题前逐一调查每个事项。
Step 2: Investigate Each Item
步骤2:调查每个事项
CRITICAL: Never create an issue without investigation. Understanding comes first.
关键注意事项:绝不跳过调查直接创建问题。必须先理解问题。
Investigation Protocol by Item Type
按事项类型划分的调查流程
For API Errors / Bugs
针对API错误/Bug
Investigation: [Item Title]
调查:[事项标题]
- Error code: [e.g., INTERNAL_ERROR, 500, 404]
- Error message: [exact message]
- Request endpoint: [URL]
- Request method: [GET/POST/etc.]
- 错误代码:[例如:INTERNAL_ERROR, 500, 404]
- 错误信息:[准确信息]
- 请求端点:[URL]
- 请求方法:[GET/POST等]
- Can reproduce: [Yes/No]
- Reproduction steps:
- [Step 1]
- [Step 2]
- 能否复现:[是/否]
- 复现步骤:
- [步骤1]
- [步骤2]
3. Code Investigation
3. 代码调查
- Relevant files: [paths]
- Likely cause: [hypothesis after code review]
- Related code: [functions/modules involved]
- 相关文件:[路径]
- 可能原因:[代码审查后的假设]
- 涉及代码:[相关函数/模块]
4. Impact Assessment
4. 影响评估
- Users affected: [All/Some/Specific conditions]
- Functionality blocked: [What can't users do?]
- Workaround exists: [Yes/No - describe if yes]
- 受影响用户:[全部/部分/特定条件下的用户]
- 被阻塞功能:[用户无法执行哪些操作?]
- 是否有替代方案:[是/否 - 若有则描述]
- Type: Bug
- Severity: [Critical/High/Medium/Low]
- Priority: [Critical/High/Medium/Low]
- 类型:Bug
- 严重程度:[Critical/高/中/低]
- 优先级:[Critical/高/中/低]
For UX/Feature Feedback
针对UX/功能反馈
Investigation: [Item Title]
调查:[事项标题]
1. Current Behavior
1. 当前行为
- What exists now: [description]
- Where it appears: [URLs/screens]
- Current implementation: [code locations]
- 现有状态:[描述]
- 出现位置:[URL/页面]
- 当前实现:[代码位置]
2. Requested Change
2. 请求的更改
- What's being asked for: [description]
- User impact: [how this affects users]
- Business context: [why this matters]
- 需求内容:[描述]
- 用户影响:[对用户的影响]
- 业务背景:[需求的重要性]
- Files affected: [list]
- Complexity: [Low/Medium/High]
- Dependencies: [other features/systems]
- 受影响文件:[列表]
- 复杂度:[低/中/高]
- 依赖项:[其他功能/系统]
4. Design Considerations
4. 设计考虑
- Options identified:
- [Option A] - [pros/cons]
- [Option B] - [pros/cons]
- Recommendation: [if clear]
- Needs: [Design input / Product decision / Research]
- 已识别的方案:
- [方案A] - [优缺点]
- [方案B] - [优缺点]
- 推荐方案:[若明确]
- 所需支持:[设计输入 / 产品决策 / 调研]
- Type: Feature / Research / UX Enhancement
- Priority: [Critical/High/Medium/Low]
- 类型:功能 / 调研 / UX优化
- 优先级:[Critical/高/中/低]
For Production Incidents
针对生产环境事故
Investigation: [Item Title]
调查:[事项标题]
1. Incident Details
1. 事故详情
- First reported: [timestamp]
- Frequency: [One-time/Intermittent/Constant]
- Environment: [Production/Staging/etc.]
- 首次上报时间:[时间戳]
- 发生频率:[一次性/间歇性/持续]
- 环境:[生产/预发布等]
- Error logs: [key log entries]
- Stack trace: [if available]
- Affected service: [component/service name]
- 错误日志:[关键日志条目]
- 堆栈跟踪:[若有]
- 受影响服务:[组件/服务名称]
3. Impact Assessment
3. 影响评估
- Users affected: [count/percentage]
- Revenue impact: [if applicable]
- SLA implications: [if applicable]
- 受影响用户:[数量/比例]
- 收入影响:[若适用]
- SLA影响:[若适用]
4. Root Cause Analysis
4. 根本原因分析
- Hypothesis: [likely cause]
- Evidence: [supporting data]
- Related changes: [recent deployments/changes]
- 假设:[可能原因]
- 证据:[支持数据]
- 相关变更:[近期部署/变更]
- Type: Bug
- Severity: Critical / High
- Priority: Critical / High
- 类型:Bug
- 严重程度:Critical / 高
- 优先级:Critical / 高
Investigation Checklist
调查检查清单
Step 3: Classify Each Item
步骤3:对每个事项进行分类
| Type | When to Use | Project Board Type |
|---|
| Bug | Something broken, not working as designed | Bug |
| Feature | New capability, clear requirements | Feature |
| Research | Needs exploration, design thinking, options analysis | Research |
| Spike | Time-boxed technical investigation | Spike |
| Chore | Maintenance, cleanup, non-user-facing | Chore |
| UX Enhancement | Improving existing user experience | Feature |
| 类型 | 使用场景 | 项目看板类型 |
|---|
| Bug | 功能损坏,未按设计工作 | Bug |
| 功能 | 新功能,需求明确 | Feature |
| 调研 | 需要探索、设计思考、方案分析 | Research |
| 技术探索 | 时间盒限制的技术调研 | Spike |
| 维护任务 | 维护、清理、非用户可见的任务 | Chore |
| UX优化 | 改进现有用户体验 | Feature |
Priority Classification
优先级分类
| Priority | Criteria | Response |
|---|
| Critical | Production down, data loss, security breach | Immediate |
| High | Major feature broken, significant user impact, blocking | This sprint |
| Medium | Feature degraded, workaround exists, important but not blocking | Next sprint |
| Low | Minor issue, cosmetic, nice-to-have | Backlog |
| 优先级 | 判定标准 | 响应要求 |
|---|
| Critical | 生产环境宕机、数据丢失、安全漏洞 | 立即处理 |
| 高 | 主要功能损坏、用户影响大、阻塞工作 | 本迭代处理 |
| 中 | 功能降级、有替代方案、重要但不阻塞 | 下一迭代处理 |
| 低 | 小问题、 cosmetic问题、锦上添花的功能 | 放入待办列表 |
Severity vs Priority
严重程度 vs 优先级
- Severity = How bad is the problem? (Technical assessment)
- Priority = How soon should we fix it? (Business decision)
A low-severity bug affecting a VIP customer may be high priority.
A high-severity bug on a deprecated feature may be low priority.
- 严重程度 = 问题的严重程度?(技术评估)
- 优先级 = 我们需要多快修复?(业务决策)
影响VIP客户的低严重程度Bug可能优先级为高。
已弃用功能上的高严重程度Bug可能优先级为低。
Step 4: Create Well-Formed Issues
步骤4:创建格式规范的问题
Issue Template: Bug
Bug问题模板
bash
gh issue create \
--title "[Bug] [Concise description of the problem]" \
--body "## Summary
[One-sentence description of the bug]
bash
gh issue create \
--title "[Bug] [问题的简洁描述]" \
--body "## 摘要
[Bug的一句话描述]
- URL: [affected URL]
- Environment: [Production/Staging/Local]
- Browser/Client: [if relevant]
- URL: [受影响的URL]
- 环境: [生产/预发布/本地]
- 浏览器/客户端: [若相关]
- [Step 1]
- [Step 2]
- [Step 3]
```
[Error message, status code, or log output]
```
Investigation Findings
调查结果
- Affected files: [list of files]
- Likely cause: [hypothesis]
- Related code: [functions/modules]
- 受影响文件: [文件列表]
- 可能原因: [假设]
- 相关代码: [函数/模块]
- Users affected: [scope]
- Functionality blocked: [what can't users do]
- Workaround: [if any]
- 受影响用户: [范围]
- 被阻塞功能: [用户无法执行的操作]
- 替代方案: [若有]
[Any additional technical context from investigation]
Source: UAT Feedback / User Report / Support Ticket
Reported: [DATE]"
[调查得到的其他技术背景]
来源: UAT反馈 / 用户报告 / 支持工单
上报时间: [日期]"
Issue Template: Feature / UX Enhancement
功能/UX优化问题模板
bash
gh issue create \
--title "[Feature] [Clear description of the feature]" \
--body "## Summary
[One-sentence description of what this feature does]
bash
gh issue create \
--title "[Feature] [功能的清晰描述]" \
--body "## 摘要
[功能的一句话描述]
[Why this is needed - user feedback, business requirement, etc.]
[What exists today, if anything]
[What should happen after implementation]
As a [type of user], I want [goal] so that [benefit].
Investigation Findings
调查结果
- Files affected: [list]
- Complexity: [Low/Medium/High]
- Dependencies: [related features/systems]
- 受影响文件: [列表]
- 复杂度: [低/中/高]
- 依赖项: [相关功能/系统]
Design Considerations
设计考虑
[Options considered, recommendations, questions]
- [What this issue will NOT address]
[Implementation hints, patterns to follow, etc.]
Source: UAT Feedback / User Request / Stakeholder Input
Requested: [DATE]"
[实现提示、遵循的模式等]
来源: UAT反馈 / 用户请求 / 利益相关者输入
请求时间: [日期]"
Issue Template: Research / Spike
调研/技术探索问题模板
bash
gh issue create \
--title "[Research] [What needs to be understood]" \
--body "## Summary
[One-sentence description of what needs to be researched]
bash
gh issue create \
--title "[Research] [需要调研的内容]" \
--body "## 摘要
[需要调研内容的一句话描述]
[Why this research is needed]
Questions to Answer
需要解答的问题
- [Question 1]
- [Question 2]
- [Question 3]
Investigation Context
调查背景
[What was discovered during initial triage that prompted this research]
[Suggested time limit: e.g., 4 hours, 1 day]
Source: UAT Feedback / Triage Investigation
Created: [DATE]"
来源: UAT反馈 / 分类处理调研
创建时间: [日期]"
Step 5: Add to Project Board (MANDATORY)
步骤5:添加至项目看板(必填)
Every issue MUST be added to the project board with correct fields.
After creating issue, add to project
After creating issue, add to project
ISSUE_URL=$(gh issue view [ISSUE_NUMBER] --json url -q '.url')
gh project item-add "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER" --url "$ISSUE_URL"
ISSUE_URL=$(gh issue view [ISSUE_NUMBER] --json url -q '.url')
gh project item-add "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER" --url "$ISSUE_URL"
ITEM_ID=$(gh project item-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r ".items[] | select(.content.number == [ISSUE_NUMBER]) | .id")
ITEM_ID=$(gh project item-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r ".items[] | select(.content.number == [ISSUE_NUMBER]) | .id")
Get field IDs
Get field IDs
PROJECT_ID=$(gh project list --owner "$GH_PROJECT_OWNER" --format json |
jq -r ".projects[] | select(.number == $GITHUB_PROJECT_NUM) | .id")
STATUS_FIELD_ID=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r '.fields[] | select(.name == "Status") | .id')
TYPE_FIELD_NAME="Type"
if ! gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER" --format json | jq -e '.fields[] | select(.name == "Type")' >/dev/null 2>&1; then
if gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER" --format json | jq -e '.fields[] | select(.name == "Issue Type")' >/dev/null 2>&1; then
TYPE_FIELD_NAME="Issue Type"
fi
fi
TYPE_FIELD_ID=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r --arg type_field "$TYPE_FIELD_NAME" '.fields[] | select(.name == $type_field) | .id')
PRIORITY_FIELD_ID=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r '.fields[] | select(.name == "Priority") | .id')
PROJECT_ID=$(gh project list --owner "$GH_PROJECT_OWNER" --format json |
jq -r ".projects[] | select(.number == $GITHUB_PROJECT_NUM) | .id")
STATUS_FIELD_ID=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r '.fields[] | select(.name == "Status") | .id')
TYPE_FIELD_NAME="Type"
if ! gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER" --format json | jq -e '.fields[] | select(.name == "Type")' >/dev/null 2>&1; then
if gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER" --format json | jq -e '.fields[] | select(.name == "Issue Type")' >/dev/null 2>&1; then
TYPE_FIELD_NAME="Issue Type"
fi
fi
TYPE_FIELD_ID=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r --arg type_field "$TYPE_FIELD_NAME" '.fields[] | select(.name == $type_field) | .id')
PRIORITY_FIELD_ID=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r '.fields[] | select(.name == "Priority") | .id')
Set Status = Ready (or Backlog for lower priority)
Set Status = Ready (or Backlog for lower priority)
READY_OPTION_ID=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r '.fields[] | select(.name == "Status") | .options[] | select(.name == "Ready") | .id')
gh project item-edit --project-id "$PROJECT_ID" --id "$ITEM_ID"
--field-id "$STATUS_FIELD_ID" --single-select-option-id "$READY_OPTION_ID"
READY_OPTION_ID=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r '.fields[] | select(.name == "Status") | .options[] | select(.name == "Ready") | .id')
gh project item-edit --project-id "$PROJECT_ID" --id "$ITEM_ID"
--field-id "$STATUS_FIELD_ID" --single-select-option-id "$READY_OPTION_ID"
Set Type (Bug, Feature, Research, etc.)
Set Type (Bug, Feature, Research, etc.)
TYPE_OPTION_ID=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r --arg type_field "$TYPE_FIELD_NAME" --arg type_value "[TYPE]" '.fields[] | select(.name == $type_field) | .options[] | select(.name == $type_value) | .id')
gh project item-edit --project-id "$PROJECT_ID" --id "$ITEM_ID"
--field-id "$TYPE_FIELD_ID" --single-select-option-id "$TYPE_OPTION_ID"
TYPE_OPTION_ID=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r --arg type_field "$TYPE_FIELD_NAME" --arg type_value "[TYPE]" '.fields[] | select(.name == $type_field) | .options[] | select(.name == $type_value) | .id')
gh project item-edit --project-id "$PROJECT_ID" --id "$ITEM_ID"
--field-id "$TYPE_FIELD_ID" --single-select-option-id "$TYPE_OPTION_ID"
PRIORITY_OPTION_ID=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r ".fields[] | select(.name == "Priority") | .options[] | select(.name == "[PRIORITY]") | .id")
gh project item-edit --project-id "$PROJECT_ID" --id "$ITEM_ID"
--field-id "$PRIORITY_FIELD_ID" --single-select-option-id "$PRIORITY_OPTION_ID"
PRIORITY_OPTION_ID=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER"
--format json | jq -r ".fields[] | select(.name == "Priority") | .options[] | select(.name == "[PRIORITY]") | .id")
gh project item-edit --project-id "$PROJECT_ID" --id "$ITEM_ID"
--field-id "$PRIORITY_FIELD_ID" --single-select-option-id "$PRIORITY_OPTION_ID"
echo "Issue #[ISSUE_NUMBER] added to project with Status=Ready, Type=[TYPE], Priority=[PRIORITY]"
**Skill:** `project-board-enforcement`
---
echo "Issue #[ISSUE_NUMBER] added to project with Status=Ready, Type=[TYPE], Priority=[PRIORITY]"
**技能依赖:** `project-board-enforcement`
---
Step 6: Summary Report
步骤6:汇总报告
After all items are triaged, provide a summary:
| # | Issue | Type | Priority | Status |
|---|
| 1 | #123 - Family page API error | Bug | High | Ready |
| 2 | #124 - Kin Circle terminology research | Research | Medium | Ready |
| 3 | #125 - Cannot add care recipient | Bug | High | Ready |
| 序号 | 问题 | 类型 | 优先级 | 状态 |
|---|
| 1 | #123 - 家庭页面API错误 | Bug | 高 | 就绪 |
| 2 | #124 - Kin Circle术语调研 | 调研 | 中 | 就绪 |
| 3 | #125 - 无法添加被照护人 | Bug | 高 | 就绪 |
Project Board Status
项目看板状态
All issues added to project board with correct fields.
- #123 - Blocking user access to family page
- #125 - Blocking care recipient management
- #124 - UX research can proceed in parallel
- #123 - 阻塞用户访问家庭页面
- #125 - 阻塞被照护人管理
- #124 - UX调研可并行进行
Example: Processing UAT Feedback
示例:处理UAT反馈
markdown
We have some bugs to resolve from UAT.
markdown
我们有一些来自UAT的bug需要修复。
Family Page has error
家庭页面出现错误
https://stage.kin.life/family
Displays: Something went wrong on our end.
[curl command and error response]
https://stage.kin.life/family
显示:Something went wrong on our end.
[curl command and error response]
Family Page refers to "Children" - not Care Recipients
家庭页面使用“Children”而非“Care Recipients”
Need better terminology, research required.
Family Page unable to add care recipient
家庭页面无法添加被照护人
[curl command and error response]
[curl command and error response]
Step 1: Parse into items
- Item 1: Family page 500 error on load
- Item 2: Terminology issue (Children vs Care Recipients)
- Item 3: Cannot add care recipient (500 error)
Step 2: Investigate each
Item 1 Investigation:
- API endpoint:
/api/v1/children?limit=50
- Error:
- Code review: Check children service, database queries
- Impact: Users cannot view family page at all
Item 2 Investigation:
- Current: Hardcoded "Children" but button says "care recipient"
- Request: More inclusive language for diverse care relationships
- Scope: UI changes, possibly data model changes
- Needs: UX research, product decision
Item 3 Investigation:
- API endpoint:
- Payload: Valid care recipient data
- Error:
- Code review: Check creation endpoint, validation, database writes
- Impact: Users cannot add any care recipients
Step 3: Classify
| Item | Type | Priority | Reasoning |
|---|
| 1 | Bug | High | Core functionality broken |
| 2 | Research | Medium | UX improvement, not blocking |
| 3 | Bug | High | Core functionality broken |
Step 4: Create issues (using templates above)
Step 5: Add to project board (with Type, Priority, Status)
Step 6: Summary report (as shown above)
步骤1:拆分为事项
- 事项1:家庭页面加载时出现500错误
- 事项2:术语问题(Children vs Care Recipients)
- 事项3:无法添加被照护人(500错误)
步骤2:逐一调查
事项1调查结果:
- API端点:
/api/v1/children?limit=50
- 错误:
- 代码审查: 检查children服务、数据库查询
- 影响: 用户完全无法查看家庭页面
事项2调查结果:
- 当前状态: 硬编码使用“Children”,但按钮显示“care recipient”
- 请求: 为多样化的照护关系使用更具包容性的术语
- 范围: UI更改,可能涉及数据模型更改
- 需求: UX调研、产品决策
事项3调查结果:
- API端点:
- 请求负载: 有效的被照护人数据
- 错误:
- 代码审查: 检查创建端点、验证逻辑、数据库写入
- 影响: 用户无法添加任何被照护人
步骤3:分类
| 事项 | 类型 | 优先级 | 理由 |
|---|
| 1 | Bug | 高 | 核心功能损坏 |
| 2 | 调研 | 中 | UX优化,不阻塞业务 |
| 3 | Bug | 高 | 核心功能损坏 |
步骤4:创建问题(使用上述模板)
步骤5:添加到项目看板(设置类型、优先级、状态)
步骤6:汇总报告(如上所示)
Best Practices for Issue Quality
问题质量最佳实践
| Type | Format | Example |
|---|
| Bug | | [Bug] Family page returns 500 error
|
| Feature | | [Feature] Add person to Kin Circle
|
| Research | [Research] <What to investigate>
| [Research] Inclusive terminology for care relationships
|
| Spike | [Spike] <Technical question>
| [Spike] Evaluate API caching strategies
|
| 类型 | 格式 | 示例 |
|---|
| Bug | | |
| 功能 | | [Feature] 添加人员到Kin Circle
|
| 调研 | | |
| 技术探索 | | |
Acceptance Criteria Quality
验收标准质量
Good criteria are:
- Specific and verifiable
- Written as checkboxes
- Focused on behavior, not implementation
- Testable
Bad:
Good:
好的标准应具备:
- 具体且可验证
- 以复选框形式呈现
- 关注行为而非实现
- 可测试
不好的示例:
好的示例:
Investigation Documentation
调查文档
Always document:
- What you found
- Where you looked
- What the likely cause is
- What the impact is
This saves time when implementation begins.
Handling Unclear Feedback
处理不明确的反馈
When Feedback is Vague
当反馈模糊时
If feedback lacks detail:
- Ask clarifying questions before creating issue
- Create Research issue to gather more information
- Document what IS known in the issue
如果反馈缺少细节:
- 在创建问题前提出澄清问题
- 创建调研问题以收集更多信息
- 在问题中记录已知的内容
Clarification Template
澄清模板
markdown
Before I create issues for this feedback, I need clarification:
1. **[Item 1]:** [Question about scope/expected behavior/reproduction]
2. **[Item 2]:** [Question about priority/business context]
Once clarified, I'll complete the investigation and create properly-formed issues.
markdown
在为这些反馈创建问题之前,我需要以下澄清:
1. **[事项1]:** [关于范围/预期行为/复现步骤的问题]
2. **[事项2]:** [关于优先级/业务背景的问题]
得到澄清后,我会完成调查并创建格式规范的问题。
When to Create Research Issues
何时创建调研问题
Create a Research issue instead of Bug/Feature when:
- Requirements are unclear
- Multiple solutions exist
- Design decisions needed
- User research required
- Technical feasibility unknown
当出现以下情况时,创建调研问题而非Bug/功能问题:
- 需求不明确
- 存在多个解决方案
- 需要做出设计决策
- 需要用户调研
- 技术可行性未知
Integration with Other Skills
与其他技能的集成
This skill flows TO:
本技能可流向:
| Skill | When |
|---|
| After issues created, to begin resolution |
| If a feedback item is too large for one issue |
| If feedback items should be grouped as epic |
| 技能 | 时机 |
|---|
| 创建问题后,开始解决问题时 |
| 若反馈事项过大,无法用单个问题覆盖时 |
| 若反馈事项应组合为史诗时 |
| Skill | For |
|---|
project-board-enforcement
| Adding issues to project board |
| Investigation patterns |
| Issue quality standards |
| 技能 | 用途 |
|---|
project-board-enforcement
| 将问题添加到项目看板 |
| 调研模式 |
| 问题质量标准 |
Store triage sessions in knowledge graph:
bash
mcp__memory__create_entities([{
"name": "Triage-[DATE]-[SOURCE]",
"entityType": "FeedbackTriage",
"observations": [
"Source: UAT / User Report / etc.",
"Date: [DATE]",
"Items received: [COUNT]",
"Issues created: #X, #Y, #Z",
"Types: [Bug: N, Feature: N, Research: N]",
"High priority: [COUNT]"
]
}])
将分类处理会话存储到知识图谱:
bash
mcp__memory__create_entities([{
"name": "Triage-[DATE]-[SOURCE]",
"entityType": "FeedbackTriage",
"observations": [
"Source: UAT / User Report / etc.",
"Date: [DATE]",
"Items received: [COUNT]",
"Issues created: #X, #Y, #Z",
"Types: [Bug: N, Feature: N, Research: N]",
"High priority: [COUNT]"
]
}])
Before Starting Triage
开始分类处理前
Gate: No issue is created without investigation. No issue is left outside the project board.
准入门槛:没有经过调查的问题不能创建。没有问题会被留在项目看板之外。
Proceeding to Resolution
进入问题解决阶段
If the user requests resolution after triage:
markdown
Issues have been created and prioritized.
**To resolve these issues:**
1. I will work through them using `issue-driven-development`
2. Starting with highest priority: #[N]
3. Each issue will follow the full development process
Shall I proceed with resolution, or should these be assigned for later work?
If proceeding, invoke
for each issue in priority order.
如果用户在分类处理后请求解决问题:
markdown
问题已创建并完成优先级排序。
**解决这些问题的方式:**
1. 我会使用`issue-driven-development`来处理它们
2. 从最高优先级的问题开始:#N
3. 每个问题都会遵循完整的开发流程
我是否应该开始解决这些问题,还是将它们分配给后续工作?