中文 | English
一套 Claude Code / Agent Skill,用盖洛普(Gallup)50 余年职场研究的循证框架,帮你做组织诊断、敬业度调查、优势识别与管理者赋能。 一个仓库,两个配套 skill,按视角分:底层框架(Q12 / CliftonStrengths / 经理人 70% 杠杆 / 循证 HR)完全共享,只是应用高度不同。 方法论顾问定位,不做角色扮演。 安装时会先问你装哪个视角(见怎么用)。
gallup-framework 企业文化部版 |
gallup-for-managers 管理者版 |
|
|---|---|---|
| 给谁 | 企业文化 / HR / OD 部门 | 一线带团队的管理者 |
| 视角 | 设计者:怎么设计、分析、理解、推动 | 操作者:我的团队该怎么带 |
| 典型场景 | 设计敬业度调查、解读数据、写汇报、设计机制、培训管理者 | 拿到团队报告怎么开对话、日常怎么带团队、怎么用下属优势 |
| 特点 | 全(含设计与分析轨 + 赋能管理者层 + 调研与证据 + 边界) | 轻(第二人称、动作优先、自包含单文件) |
一句话:做设计和分析,用文化部版;带自己的团队,用管理者版。 文化部版是超集,管理者版是为操作者重写的精简独立版。安装器会让你二选一(或都装)。
- 本套件是对盖洛普公开方法论的框架提炼 + 有限引用,与 Gallup, Inc. 无任何隶属或授权关系。Gallup®、Q12®、CliftonStrengths®、StrengthsFinder® 等为其各自所有者的商标。
- 它提供框架视角,不能替代正版 Q12 / CliftonStrengths 测评,也无法生成你公司的真实分数。
- 引用的效果数字是盖洛普公开发布的、卖方立场的研究,当方向性参考,非铁证;注意相关 ≠ 因果。
- 不可用作招聘录用 / 淘汰、绩效奖惩、心理评估的直接依据。
- 证据以西方职场为主,迁移到中国等组织语境需本地化校准,不保证跨文化适用。
① 测量敬业度(Q12) → ② 识别天赋(CliftonStrengths) → ③ 经理人据此带人(70%杠杆) → ④ 循证闭环(测-比-干预-再测)
↑________________________________ 再测量 ________________________________↓
- Q12:12 道题测团队敬业度,分四层金字塔(基本需求 → 个人贡献 → 团队归属 → 共同成长)。
- CliftonStrengths:34 个天赋主题、四大领域,核心是"扬长而非补短"。
- 经理人杠杆:经理人解释了团队敬业度约 70% 的方差,所以团队问题的第一诊断对象是经理。
- 循证纪律:人事判断要挂在数据和对标上,而非凭主观臆断。
两个 skill 都内置交互式诊断:遇到"XX 题分低说明什么 / 该怎么做",先用框架讲清要害,再反问你 1~3 句精准定位,根据你的回答给针对性动作,像顾问陪你诊断,而不是直接抛出一份标准清单。
你:我们刚做完敬业度调查,研发部 Q04(过去七天有没有被认可)分特别低,我该怎么跟领导汇报、怎么办?
它:
先给框架初判:Q04 在金字塔 L2「个人贡献」层,测即时认可。低分通常是三种情况:
- 认可缺失 / 不及时;
- 经理给反馈的方式有问题(认可主要是经理的动作,紧贴 70% 杠杆);
- 认可有,但员工感知不到。
但要给你能下手的方案,我得先确认:
- 整个研发部普遍低,还是个别小组?(全研发 → 机制 / 文化问题;个别组 → 那个经理)
- 同在 L2 的 Q03 / Q05 / Q06 分怎样?(一起低 = 整层坍塌;单 Q04 低 = 认可机制缺失,相对好解决)
- 你们现在有没有认可机制?
(据回答给方案,如"全研发、单 Q04 低、无机制"→)
- 汇报:别只报分数,报"卡在 L2 认可环 + 建议",用"认可缺口直接压敬业度、而认可主要是经理动作"组织论点。
- 怎么办:设计轻量即时认可机制 + 培训经理"具体化认可",而非做一次性的年度表彰。
- 提醒:引盖洛普数字标一句"卖方研究、方向性参考"。
你:我手下有个人最近状态一直很低迷,怎么办?
它:
先别急着贴"态度问题"标签。在盖洛普框架里,这通常是金字塔某一层坍塌了,而你这个直接经理是最大的变量(70% 在你手上)。常见三种:
- L1:他不清楚你要什么;
- L2:没被认可,或没用上他擅长的事;
- L4:看不到成长。
给动作前我先问三句:
- 是他一个人这样,还是团队普遍?
- 你上次单独和他聊是什么时候?
- 他最擅长的事,现在的工作用得上吗?
(据回答给动作,如"就他、很久没单聊、强项没用上"→)
- 本周约他做一次 15 分钟快速连接,别谈 KPI,问"最近有什么挡路的、什么让你最有干劲";
- 看能不能调一块发挥他强项的任务给他;
- 下次签到再回看。
- 企业文化 / HR / OD / HRBP:做敬业度调查、组织诊断、给管理层汇报、设计人才机制、培训管理者 → 用
gallup-framework。 - 一线管理者 / 新晋经理 / 团队负责人:带好自己的团队 → 用
gallup-for-managers。
推荐:用安装器,它会先问你装哪个视角再装。
git clone https://github.com/zc790-hub/gallup-hr-skills.git
cd gallup-hr-skills
bash install.sh
# 按提示选:1) 文化部版 2) 管理者版 3) 两个都装安装器默认装到 Claude Code 的 ~/.claude/skills/;装到其他 Agent 用 SKILLS_DIR=/你的/skills目录 bash install.sh。
或手动复制你需要的那个子目录:
cp -r gallup-framework ~/.claude/skills/ # 文化部 / HR 用
cp -r gallup-for-managers ~/.claude/skills/ # 管理者用之后在对话里描述你的组织 / 团队问题,skill 会自动触发。
gallup-hr-skills/
├── install.sh # 安装器(先问装哪个视角再装)
├── gallup-framework/ # 企业文化部版(超集)
│ ├── SKILL.md
│ └── references/
│ ├── framework-extraction.md # 框架提炼蓝图
│ └── research/ # 四篇调研(Q12 / 优势 / 经理人 / 证据与批评)
└── gallup-for-managers/ # 管理者精简版(自包含单文件)
└── SKILL.md
见上方免责声明。核心原则:这是帮你思考、设计、产出的循证框架顾问,不是数据系统,也不是评判人的标尺。盖洛普方法论本身也有学界争议(敬业度构念模糊、相关≠因果、商业利益冲突、CliftonStrengths 部分信效度软肋),两个 skill 内部都如实标注。
- 框架蒸馏自 Gallup 公开发布的研究与著作(First, Break All the Rules、Q12 Meta-Analysis、State of the Global Workplace、CliftonStrengths 等),版权归原作者 / Gallup, Inc. 所有。
- 学界批评部分参考 Macey & Schneider (2008) 等公开文献。
- 本套件由 女娲 · Skill造人术 生成,创建者 花叔。
MIT © 2026 zc790-hub
A Claude Code / Agent Skill that uses Gallup's evidence-based HR framework (50+ years of workplace research) to help you run organizational diagnosis, engagement surveys, strengths identification, and manager enablement. One repo, two companion skills split by perspective. They share the same core framework (Q12 / CliftonStrengths / the 70% manager lever / evidence-based HR); only the altitude of application differs. Methodology advisor, not roleplay. The installer asks which perspective you want before installing (see How to use).
gallup-framework Culture/HR edition |
gallup-for-managers Manager edition |
|
|---|---|---|
| For whom | Corporate culture / HR / OD teams | Frontline people managers |
| Perspective | Designer: how to design, analyze, understand, drive | Operator: how do I lead my own team |
| Typical use | Design engagement surveys, interpret data, brief leadership, design mechanisms, train managers | Run the post-survey conversation, lead day to day, use reports' strengths |
| Shape | Full (design & analysis track + manager-enablement layer + research & evidence + boundaries) | Lean (second-person, action-first, self-contained single file) |
In one line: for design & analysis, use the Culture/HR edition; for leading your own team, use the Manager edition. The Culture edition is the superset; the Manager edition is a lean standalone rewritten for operators. The installer lets you pick one (or both).
- This toolkit is a framework distillation + limited citation of Gallup's publicly available methodology. It is not affiliated with or endorsed by Gallup, Inc. Gallup®, Q12®, CliftonStrengths®, StrengthsFinder® are trademarks of their respective owners.
- It offers a framework lens and cannot replace the official Q12 / CliftonStrengths assessments, nor generate your company's real scores.
- Cited effect sizes are Gallup's own, vendor-side published research; treat them as directional, not proof, and note that correlation does not equal causation.
- Do not use as a direct basis for hiring/firing, performance rewards/penalties, or psychological screening.
- Evidence is Western-workplace-heavy; applying it to other organizational/cultural contexts needs local calibration.
① Measure engagement (Q12) → ② Identify talent (CliftonStrengths) → ③ Managers lead on it (70% lever) → ④ Evidence loop (measure-benchmark-act-remeasure)
- Q12: 12 items measuring team engagement, in a four-level pyramid (Basic Needs → Individual → Teamwork → Growth).
- CliftonStrengths: 34 talent themes across four domains; "invest in strengths, manage weaknesses."
- Manager lever: managers explain ~70% of the variance in team engagement, so diagnose the manager first.
- Evidence discipline: anchor people decisions in measurement and benchmarks, not gut feel.
Both skills feature interactive diagnosis: for "what does a low score on X mean / what do I do," they first frame it, then ask you 1 to 3 targeted questions, then give context-specific actions. It works like a consultant diagnosing with you, not a canned answer.
Culture/HR edition (designer):
You: We just ran an engagement survey; R&D scored very low on Q04 (recognition in the last 7 days). How do I brief leadership and what do I do?
It:
Framing first: Q04 sits at pyramid L2 "Individual", measuring timely recognition. A low score usually means one of:
- recognition is missing or untimely;
- the manager's feedback habits are off (recognition is mostly a manager action, the 70% lever);
- it exists but isn't felt.
To give you something actionable I need:
- whole R&D, or one team? (all → mechanism/culture; one → that manager)
- how do neighboring L2 items Q03/Q05/Q06 score? (all low = whole layer collapsed; only Q04 = recognition gap)
- any recognition mechanism today?
(then a targeted plan, e.g. "all R&D, only Q04 low, no mechanism" →)
- Brief: don't just report the score; report "stuck at the L2 recognition link + recommendation."
- What to do: design a lightweight timely-recognition mechanism + coach managers on specific recognition, not a one-off annual award.
- Caveat: flag vendor-data caveats when citing Gallup numbers.
Manager edition (operator):
You: One of my reports has seemed checked-out lately. What do I do?
It:
Don't slap an "attitude problem" label on it yet. In Gallup's frame this is usually a collapsed pyramid layer, and you, the direct manager, are the biggest variable (70% is on you). Three common causes:
- L1: he's unclear what you expect;
- L2: unrecognized, or not using his strengths;
- L4: sees no growth.
Before actions, three questions:
- just him, or the whole team?
- when did you last talk one-on-one?
- are his strengths used in his current work?
(then actions, e.g. "just him, long since a 1:1, strengths unused" →)
- a 15-minute Quick Connect this week, no KPIs, just "what's blocking you, what energizes you";
- see if you can shift a strengths-fit task to him;
- revisit at the next check-in.
- Culture / HR / OD / HRBP: surveys, diagnosis, leadership briefs, talent mechanisms, manager training →
gallup-framework. - Frontline / new managers / team leads: lead your own team well →
gallup-for-managers.
Recommended: the installer asks which perspective you want, then installs it.
git clone https://github.com/zc790-hub/gallup-hr-skills.git
cd gallup-hr-skills
bash install.sh
# choose: 1) Culture/HR 2) Manager 3) bothIt defaults to Claude Code's ~/.claude/skills/; for other agents use SKILLS_DIR=/your/skills/dir bash install.sh.
Or copy manually:
cp -r gallup-framework ~/.claude/skills/ # Culture / HR
cp -r gallup-for-managers ~/.claude/skills/ # ManagersThen describe your organization/team problem in chat and the skill triggers automatically.
See the Disclaimer. Core principle: this is an evidence-based framework advisor for thinking, designing, and producing, not a data system and not a yardstick for judging people. Gallup's methodology itself has academic critiques (construct ambiguity of "engagement," correlation versus causation, commercial conflict of interest, some CliftonStrengths reliability gaps); both skills flag these honestly.
- Distilled from Gallup's publicly published research and books (First, Break All the Rules, the Q12 Meta-Analysis, State of the Global Workplace, CliftonStrengths, etc.); copyright belongs to the original authors / Gallup, Inc.
- Academic critique draws on public literature incl. Macey & Schneider (2008).
- Generated with Nüwa · Skill Creator; creator Huashu.
MIT © 2026 zc790-hub