AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Agf Writing Github Issue

skill-pcliangx-appgenesisforge-agf-writing-github-issue · by pcliangx

Use whenever a user, product-lead, or qa-engineer wants to create a GitHub issue in the project repo — including phrases like "提一个 issue / 写一个 issue / 报 bug / 把这个开成 issue / gh issue / 上 GitHub / track 一下 / 立个 ticket". Provides the required-field skeleton, locked label set (type / area / epic / priority / severity / phase), gh CLI heredoc template, and the QA-auto-issue exception path. Replaces ad…

No reviews yet
0 installs
28 views
0.0% view→install

Install

$ agentstack add skill-pcliangx-appgenesisforge-agf-writing-github-issue

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-pcliangx-appgenesisforge-agf-writing-github-issue)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
2mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Agf Writing Github Issue? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Writing a GitHub Issue

Use this skill when any of the following:

  • 用户说 "写 issue / 提一个 issue / 报 bug / 上 GitHub / track 一下 / 立个 ticket"
  • product-lead 把 PRD 里的 AC 拆成可分派的 issue
  • dev 在 SIT 自跑中发现 P0 / P1 缺陷(特殊路径,见下
  • qa-engineer 在 E2E / UAT 中发现 P0 / P1 问题(特殊路径,见下
  • 主 Claude / 任意 agent 准备调用 gh issue create

> 拼 body / 写命令前必读 [references/templates-and-examples.md](./references/templates-and-examples.md):Body 模板(feature + bug 增补段)、gh CLI HEREDOC 正确写法、3 个完整例子(feature / bug / dev SIT 自动路径,含 RolexOps 实例需替换为本项目实际值)。

最小输入模式(用户授权 / 2026-05-13)

> 默认行为:用户只给"关键信息",agent 智能补全其他字段,直接 gh issue create,不走草稿 gate。

用户必须给的最少信息

| 信息 | 说明 | |---|---| | 要解决什么 / 想做什么 | 一两句话描述问题或功能想法,bug / feature / chore 可推断 |

就这一条——其他字段都由 agent 推断 + 补全。

Agent 必须自动补全的字段

| 字段 | 推断规则 | |---|---| | Title | 按用户描述拟动宾结构 + type prefix feat(...) / fix(...) / chore(...) / docs(...),≤70 char | | Why / 背景 | 从用户原话扩写 1-2 段;若有相关 PRD / ADR / 历史 issue,主动 grep docs/prd/ docs/adr/ 并 cite | | AC | 按 feature/bug 套模板生成 2-4 条可验证条件;bug 必含 "修复后用什么验证" | | type label | 按关键词判 feat(新功能)/ bug(修缺陷)/ chore(重构/构建/文档)/ adr(架构决策) | | area label | grep 用户描述关键词:frontend/React/UI/组件 → area:frontend;API/endpoint/后端/数据库/migration → area:backend;LLM/prompt/豆包/Qwen/embedding → area:ai;docker/compose/caddy/redis → area:infra | | priority | bug 默认 P1(核心流程影响);feature 默认 P2(用户说"急/紧急/blocking"→ P1);data loss / 安全 / 线上挂 → P0 | | severity(bug 专用) | 与 priority 同步:P0/P1 | | epic 关联 | grep 当前 branch 名(release/v1.6.0 etc.)+ 最近 commit + docs/prd/,能锁定 Epic N 才加;不确定不加,宁缺勿乱 | | phase | 默认不加;除非用户在 SIT/E2E/UAT 上下文中报问题 | | 复现步骤(bug 专用) | 用户没给则合理推测填,末尾标 ` | | **环境**(bug 专用) | grep docker compose ps` 拿当前 image / commit hash,前端 bug 默认填 "Chrome 最新版(用户未指定)" | | 关联 (Refs) | 主动 grep 关联 PRD / ADR / 现有 issue(gh issue list -L 10 关键词);找到的填,找不到填 N/A |

兜底问号(只在以下情况问用户)

  1. priority 真拿不准 P0/P1/P2 — 如用户说"挺重要",bug 但没说阻塞与否
  2. area 完全猜不出 — 描述太抽象("那个东西坏了")
  3. 是 bug 还是 feature 真分不清 — 如"X 行为不太对,是不是应该 Y"

非这三种 → 不要问,直接补全 + 创建(事后用户在 GitHub 上 edit 比中途问还快)。

自动 issue 路径(已授权,跳过用户确认)

路径 A — dev 在 SIT 中发现 P0 / P1(SIT 由 dev 自跑):

  1. 直接 gh issue create(不需问用户、不需 product-lead 同意)
  2. --label "type:bug,priority:P0,severity:P0,phase:sit"(或对应 P1)
  3. 创建后通过 SendMessage 向 product-lead 报告 issue 号
  4. P2 问题不开 issue,只在 progress/.md**SIT 证据** 段记 fail/blocked

路径 B — qa-engineer 在 E2E / UAT 中发现 P0 / P1

  1. 直接 gh issue create
  2. --label "type:bug,priority:P0,severity:P0,phase:e2e"(或 phase:uat
  3. 创建后通过 SendMessage 向 product-lead 报告 issue 号
  4. P2 问题不开 issue,只记录在 E2E / UAT 报告里

来源:用户 2026-05-13 授权;SIT 归 dev,QA 仅覆盖 E2E / UAT。

优先级判定

| 级别 | 适用场景 | |---|---| | P0 | 线上阻塞 / 数据损坏 / 安全漏洞 / UAT-block regression | | P1 | 核心流程影响 / UX regression / 性能 SLA 超标但非阻塞 | | P2 | 次要 / 体验优化 / nice-to-have |

判不准时 → 偏保守降一级(P0→P1),勿默认拔高。

本仓锁定 label 集合(不要自创)

gh label list 已固化的 label,只能从下面选(新需求要新 label → 先停下问用户,不要直接 gh label create):

type:feat | type:bug | type:chore | type:adr
area:frontend | area:backend | area:ai | area:infra
epic:1 | epic:2 | epic:3 | epic:4 | epic:5 | epic:6
epic:7 | epic:8 | epic:9 | epic:10 | epic:11 | epic:12
phase:design | phase:dev | phase:review | phase:sit | phase:e2e | phase:uat
priority:P0 | priority:P1 | priority:P2
severity:P0 | severity:P1
status:blocked | status:needs-info | status:wontfix
release:vX.Y   # release 关联标签,按项目实际 milestone 替换(模板无固定值)

gh CLI 写法红线

正确写法 = HEREDOC --body "$(cat <<'EOF' ... EOF)"(完整模板见 [references/templates-and-examples.md](./references/templates-and-examples.md))。错误做法(禁止):

  • gh issue create --body "..." 含换行 / 反引号 / 中文引号 —— shell 转义噩梦
  • --label "P0" —— 不带 priority: 前缀
  • gh label create 自创新 label —— 必须先问用户

执行流程(最小输入模式)

  1. 接收「关键信息」→ 按上文"Agent 必须自动补全的字段"表推断 title / type / area / priority / severity + grep epic / refs,按 type 套 body 模板(见 references)。
  2. 命中"兜底问号"三类才简短问一句(且只问那一个不确定字段),否则跳过。
  3. HEREDOC 拼 body + 锁定 label → gh issue create(直接执行)→ 报告 URL + label 清单 →(QA 路径)SendMessage product-lead 报 issue #。

验证 gate(issue 创建后)

每次 gh issue create 成功后,最后一条 user-facing 消息必须包含

  • ✅ Issue URL(gh 命令的 stdout 第一行)
  • ✅ 已贴的 label 清单
  • ✅(若 QA 路径)已发给 product-lead 的 SendMessage 确认

让用户能立刻点链接 + 确认分类对不对。

Anti-patterns

| 反模式 | 为什么不行 | |---|---| | Title 用句号结尾 / 含特殊字符 | gh search / 看板列表不易扫读 | | AC 写 "正常工作" / "符合预期" | 不可验证 → code-reviewer / qa-engineer 没法逐条核 | | Body 大段无结构 | 后续被分派的 dev 找不到关键信息 | | 不标 priority | 看板按 priority 排序 / 触发不了 P0 通知机制 | | 既不标 type 也不标 area | 责任田不清 → 没人接 | | Mock / fake 描述 | 禁 mock 描述,issue 必须真实场景 |

Related skills

  • agf-writing-prd — issue 通常源于 PRD AC,反向链回 PRD 路径
  • agf-writing-adr — 涉及架构变更时先开 ADR,再开 issue 引用
  • agf-running-sit-tests(dev-owned SIT)/ agf-writing-qa-report(E2E/UAT 报告,SIT 不在范围内) — QA 与 dev SIT 路径上下文

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.