关于 single-point-rubric-designer
为教育工作者设计单点量规的技能,以精确标准和具体证据支持学生自评、同伴反馈及教师形成性反馈。提供三栏、两栏和学生自拟标准等模式,不生成分数或等级。
主要功能
- 01设计精确、可观察的评价标准
- 02生成三栏或两栏量规
- 03支持学生自拟标准
- 04用于自评、同伴及教师反馈
- 05聚焦形成性证据而非评分
适用场景
- 01为学生提交作业前的自评设计量规
- 02为草稿阶段的同伴反馈提供框架
- 03引导学生围绕学习目标进行反思
安装
复制下面的命令,在终端中运行。
npx skills add https:/ / github.com/ garethmanning/ education-agent-skills --skill single-point-rubric-designer安装后在 Claude Code 中输入/single-point-rubric-designer确认已加载。
详细介绍
单点量规设计师
此技能的作用
此技能用于设计单点量规:一种聚焦于单一标准的工具,不设评分等级,而是留出空间填写具体证据。分析型量规会依据多个标准和多个等级进行评估;与之不同,单点量规只呈现一条精确描述质量要求的陈述——即标准——并将两侧栏留空,以便填写特定情境下的证据。
此技能可搭配或不搭配发展阶段分级系统使用。任何有学习目标和使用目的的教育工作者都可以使用它。
此技能提供两种格式:
三栏格式(默认):
| 有待提升之处 | 标准 | 优势所在 |
|---|---|---|
| 朝着达到标准的方向取得进展的具体证据 | [标准陈述] | 达到或超出标准的具体证据 |
中间栏包含标准陈述。学生或教师在左右两栏中填写具体、明确的证据。此格式适用于学生自我评估、同伴反馈和教师形成性反馈。
两栏格式(适用于年龄较小的学习者、A–B 阶段,或有特别说明时):
| 标准 | 我的反思 |
|---|---|
| [标准陈述] | 描述一个具体时刻或例子,说明你目前在这方面的情况。 |
左栏包含标准。学生在右栏中填写内容。如果采用三栏格式的双向结构超出当前阶段的实际需要,那么此格式更为合适。
**学生自拟版本:**指定 student_generated 选项时,此技能会生成一条脚手架式提示词,引导学生撰写自己的标准陈述。当目标是让学生形成自己的质量观,而不是朝着外部设定的标准努力时,就会使用此模式。
关键设计原则是标准的精确性:单点量规的价值在于填入各栏的证据具体明确,而不在于标准陈述本身。如果某项标准引出的回答很模糊(如“我正在进步”/“我做得不错”),就说明设计失败。标准必须足够精确,让学生和教师都能为每一栏找出具体、明确的证据。
此技能不会给出分数或等级。单点量规属于形成性工具。填写完成后的各栏内容是用于交流、反思和规划后续步骤的证据,而非总结性评分的依据。
依据基础
Fluckiger(2010)专门将单点量规作为一种负责任的学生自评工具提出——他认为,多等级量规会细致地描述平庸表现,而单点量规则聚焦于真正的高质量表现,并为学生提供展示自身证据的空间。Sadler(1989)确立了任何自评工具要发挥作用所需的基础条件:学生必须具备评估专长——能够在产出或评估质量之前识别质量。单点量规假定这一条件成立,并为培养这种能力提供支架:评价标准就是质量陈述,而各栏则要求学生证明自己能够在自己的作品中识别高质量证据。Andrade(2000、2013)证明,当学生参与制定评价标准时,自评最能稳定地提升学习成果——此技能中的学生自主制定模式正是对这一发现的直接落实。Boud(1995)认为,自评本身就是一项核心学术技能,而非补充活动:依据标准判断自己的作品,是学习者实现独立自主的关键。单点量规为培养这项能力提供了低门槛的切入点。Panadero 和 Jonsson(2013)回顾了关于量规的形成性评价研究,并发现,当学生使用量规进行自我调节时,量规能够促进学习;若主要将其用作评分工具,则不会。单点格式在结构上适用于自我调节,因为它要求学生生成证据,而不是选择等级。
输入模式
教育工作者必须提供:
- 学习目标:“我能……”陈述、能力描述或评估重点。例如:“我能构建论证,提出清晰的主张、提供支持性证据,并承认不同观点”/“我能在感到沮丧或不堪重负时调节自己的反应”/“我能设计一项变量控制明确的调查”
- **用途:**谁会使用量规,以及用途是什么。例如:“提交前的学生自评”/“起草期间的同伴反馈”/“教师形成性反馈”/“任务前规划,以明确质量要求”/“个人反思”
可选项(如有,则由上下文引擎注入):
- **阶段:**发展阶段(A–F)——为 A 或 B 触发双栏格式
- 格式:“three-column”或“two-column”——覆盖基于阶段的默认设置
- 由学生制定:“yes”——触发学生自主制定模式
- **任务描述:**学生要完成的作品——用于校准评价标准的精确度
- **学生水平:**年龄或年级——未提供阶段时,用于校准语言
提示词
You are designing a single-point rubric using the principles established by Fluckiger (2010) and Sadler (1989). Follow these rules PRECISELY — they are the quality standard for every rubric produced.
CRITICAL DESIGN PRINCIPLE:
Criterion precision determines evidence quality. A vague criterion produces vague column entries and the rubric fails. Write the criterion first — what does meeting this learning target actually look like as a specific, observable quality? Every other decision follows from that.
---
STEP 1 — FORMAT DECISION:
Apply this logic in order:
- If student_generated is "yes": skip to the STUDENT-GENERATED MODE section below. Do not produce a criterion.
- If format is explicitly provided as "two-column": use two-column format.
- If format is explicitly provided as "three-column": use three-column format.
- If band is A or B: use two-column format.
- If student_level suggests ages 5–9 (e.g. "Year 1", "age 6", "Kindergarten"): use two-column format.
- Default: use three-column format. Calibrate language to student_level if provided; otherwise write for a general secondary audience.
---
STEP 2 — CRITERION WRITING RULES (non-negotiable):
The criterion is a single precise statement of what quality looks like. Apply every rule:
1. **One integrated quality, not a checklist.** Do not list components. "A claim that is supported by evidence and addresses a counterargument" is one integrated quality. "Has a claim, has evidence, has a counterargument" is a checklist — it fails.
2. **Descriptive language, not evaluative language.** Describe what the work or behaviour looks like. Never use: good, excellent, thorough, adequate, appropriate, strong, clear (without specifying what clarity means). Instead: specify the observable feature.
3. **Specific enough to generate concrete evidence.** A student reading this criterion must be able to point to a specific moment, sentence, example, or action that belongs in each column. If they could only write "I did well" or "I need to improve," the criterion is too vague.
4. **Calibrated to the audience.** Neither too simple (patronising) nor too demanding (inaccessible). If band is provided, calibrate to band expectations. If student_level is provided, calibrate language to that level.
5. **Scoped to the learning target.** Does not add requirements the learning target does not contain. Does not drop requirements it does contain.
6. **Single sentence where possible.** Never more than two sentences.
Write the criterion BEFORE constructing the rubric table. State it explicitly and apply the quality check:
- Is it one integrated quality, not a list?
- Is the language descriptive not evaluative?
- Would a student know exactly what kind of evidence to write in each column?
- Is it calibrated to the audience?
- Does it match what the learning target requires, no more and no less?
If it fails any check, rewrite before continuing.
---
STEP 3 — RUBRIC OUTPUT:
**Three-column format:**
Present the rubric as a markdown table. The criterion goes in the centre column header. The outer columns contain the guidance text as a standing instruction for whoever completes the rubric — these are NOT filled in by you.
| Areas for growth | [Criterion statement] | Areas of strength |
|---|---|---|
| Record specific evidence here — moments, examples, or observations that show progress toward the criterion but do not yet fully meet it | [Criterion statement repeated] | Record specific evidence here — moments, examples, or observations that demonstrate or go beyond the criterion |
Do not fill in the evidence columns. They are left blank for the user to complete.
**Two-column format:**
| [Criterion statement] | My reflection |
|---|---|
| [Criterion statement repeated] | Describe a specific moment or example that shows where you are with this right now. |
---
STUDENT-GENERATED MODE:
When student_generated is "yes", do not produce a criterion statement. Instead produce:
**Framing sentence for the student:**
"Think about what doing this really well looks like for you personally. Write one sentence that describes it."
**Scaffolding prompts (provide 2–3):**
- "What would someone watching you be able to see or hear?"
- "What is the difference between doing this well and doing it adequately — what's the specific thing that changes?"
- "If you were helping a friend get better at this, what would you tell them to focus on?"
**Note to teacher:**
The student-generated criterion is personal and specific to this student. Do not correct, standardise, or replace it with your own language. The value is in the student's own articulation of quality — even if it is imperfect. With younger learners (Band A–B or ages 5–9), co-generate the criterion through conversation rather than expecting independent writing.
---
STEP 4 — USAGE GUIDANCE:
Identify which of the following purposes applies, apply the closest match if exact match is not found, and note the adaptation.
- **Student self-assessment before submission:** Student completes both outer columns independently using their own work as evidence. They then decide whether to revise based on what they wrote. The teacher does not complete the rubric — the student's evidence is the formative data.
- **Peer feedback:** The peer completes the columns based on their reading of the work. The owner reads the completed rubric and decides what to act on. The peer does not give a grade or level — only evidence.
- **Teacher formative feedback:** Teacher completes the columns with specific observations from the work. Shares with student as the basis for a feedback conversation, not as a verdict or grade. The evidence in the columns IS the feedback.
- **Pre-task planning:** Student completes only the "areas of strength" column as a planning tool — what specific evidence do I want to be able to put here by the end? The left column remains blank until after the task.
- **Personal reflection:** Student completes the rubric in relation to their own experience, behaviour, or growth over time — not a specific product. Evidence is drawn from memory, observation, or journal entries.
---
STEP 5 — CRITERION DESIGN NOTES:
Explain:
- Why the criterion was written this way
- What it deliberately includes and what it deliberately excludes
- What made it difficult to write precisely, if anything
---
STEP 6 — KNOWN LIMITATIONS:
Apply these standing limitations, plus any specific to the learning target provided:
1. The criterion statement requires judgment to write well — too broad produces vague evidence, too narrow becomes a checklist. If the output criterion feels like either of these, flag it to the user.
2. Single-point rubrics work best when students have already been introduced to the learning target and have some experience with the domain — they are formative tools, not introductory ones.
3. The student-generated version is only as useful as the quality of the student's self-knowledge. With younger learners the teacher may need to co-generate the criterion through conversation.
4. This skill does not produce grades or summative levels. If a summative judgment is needed, use coherent-rubric-logic-builder or criterion-referenced-rubric-generator instead.
---
Your task is to design a single-point rubric for:
**Learning target:** {{learning_target}}
**Purpose:** {{purpose}}
The following optional context may or may not be provided. Use whatever is available; ignore any fields marked "not provided."
**Band:** {{band}} — if not provided, use three-column format unless student_level indicates ages 5–9.
**Format:** {{format}} — if provided, overrides band-based default.
**Student generated:** {{student_generated}} — if "yes", produce the student-generated prompt instead of a criterion.
**Task description:** {{task_description}} — if provided, use to calibrate criterion precision.
**Student level:** {{student_level}} — if provided, use to calibrate language.
Return your output in this exact format:
## Single-Point Rubric: [Learning Target Summary]
**Learning target:** [The exact learning target as provided]
**Purpose:** [The stated purpose]
**Format selected:** [Three-column / Two-column / Student-generated] — [brief reason]
### Criterion
[State the criterion here before the table. One sentence, two maximum.]
**Quality check:**
- One integrated quality, not a list: [Yes / No — if No, explain and rewrite]
- Descriptive not evaluative: [Yes / No — if No, explain and rewrite]
- Generates concrete evidence: [Yes / No — if No, explain and rewrite]
- Calibrated to audience: [Yes / brief note]
- Scoped to learning target: [Yes / brief note]
### Rubric
[Insert the appropriate format table here]
### Criterion Design Notes
[Why this criterion; what it includes and excludes; what was difficult to make precise]
### Usage Guidance
[How to use this rubric for the specified purpose — what the student, peer, or teacher does with the completed columns]
### Known Limitations
[Standing limitations plus any specific to this learning target]
输出示例
示例 1
情境: 学习目标:“我能构建论证,提出清晰的主张、提供支持性证据,并承认不同观点”/用途:“提交前的学生自评”/未指定阶段/格式:三栏(默认)
单点量规:理性论证
**学习目标:**我能构建论证,提出清晰的主张、提供支持性证据,并承认不同观点 **用途:**提交前的学生自评 **所选格式:**三栏——未指定阶段,采用默认设置
标准
论证提出明确的立场,用相关证据支持该立场,并回应相反观点,解释为什么该立场依然成立。
质量检查:
- 是一个整合的质量标准,而非列表:是——这三个要素被表述为一个整合论证的组成部分,而非彼此独立的检查项
- 描述性而非评价性:是——“提出明确的立场”“用相关证据支持该立场”“回应相反观点并加以解释”都是可观察的行为;没有使用评价性语言
- 能生成具体证据:是——学生可以指出一句具体的陈述作为主张、若干具体句子或事实作为证据,以及一个具体段落作为回应反方观点的内容
- 与受众相匹配:面向普通中学生——语言通俗易懂,但不失简洁
- 与学习目标范围相符:是——该标准涵盖学习目标的全部三个组成部分,且没有超出其范围
评分量规
| 有待提升之处 | 论证提出明确的立场,用相关证据支持该立场,并回应相反观点,解释为什么该立场依然成立。 | 优势 |
|---|---|---|
| 在此记录具体证据——引用一句话或指出作品中的某个部分,说明你正在朝此目标努力,但尚未完全达标 | 论证提出明确的立场,用相关证据支持该立场,并回应相反观点,解释为什么该立场依然成立。 | 在此记录具体证据——引用一句话或指出作品中的某个部分,说明你达到了该标准或表现超出该标准 |
标准设计说明
该标准将学习目标的三个组成部分——主张、证据和反方观点——整合为对论证整体作用的单一描述。分别编写三个标准会更容易,但包含三个标准的单点量规,实际上只是一个包含三个标准的分析型量规。这种整合是有意为之:在真实论证中,这三个组成部分并非彼此独立。证据只有与主张相关联时才发挥作用;反方观点也只有与所辩护的立场相关联时才有意义。
“回应相反观点,解释为什么该立场依然成立”比“承认存在不同观点”更精确。仅仅承认可能只是一句敷衍的话(“有些人不同意”)。该标准要求学生对反方观点有所回应——在面对反方观点时坚持自己的立场。这正是学习目标的含义,也是该标准所体现的内容。
该标准不包括不属于此学习目标的写作技巧,例如段落结构、过渡和词汇。它也不涉及证据的质量(力度、来源、相关性)——相比学习目标在这一阶段的要求,这属于更高阶的期望。
使用指南
提交前的学生自我评估:
学生独立完成这份量规,并以自己的草稿作为依据。步骤如下:
- 仔细阅读评价标准,确保理解其要求。
- 结合评价标准阅读自己的草稿。
- 在右侧栏(“优势之处”)中,引用符合评价标准的具体句子,或指出作品中的具体章节。务必具体——“我的第二段”不够具体;“以‘尽管有人认为……’开头的句子”才算具体。
- 在左侧栏(“成长空间”)中,引用具体句子或章节,说明你正在朝着评价标准努力,但尚未完全达到要求。如果你认为自己已经完全达到要求,请写下要如何做才能更进一步。
- 阅读两栏内容并作出决定:提交前是否要修改?如果要,具体会做哪些修改?
教师不填写这份量规。学生提供的依据就是形成性评估数据。如果两栏中的填写内容很笼统(“我做得还行”/“我需要改进论证”),说明学生尚未理解评价标准——请先进行简短交流,再让学生修改。
已知局限
- 此评价标准面向一般中学学生。如果学生尚未接触过反方论证的概念,右侧栏中的填写内容就会很笼统——“我回应了另一方的观点”,但没有任何具体内容。分发这份量规前,应先讲解反方论证这一概念。
- 此评价标准不评估证据的质量(证据是否有力、是否相关、是否来自可信来源)。如果这些本身就是学习目标,则应使用单独的量规或评价标准。
- 提交前自我评估的效果取决于学生是否有时间据此采取行动。截止时间前 20 分钟才发放这份量规,就失去了评估的意义。自我评估后至少应安排一轮修改。
- 此技能不会生成成绩或总结性评定等级。如需总结性评价,请使用 coherent-rubric-logic-builder 或 criterion-referenced-rubric-generator。
示例 2
情境: 学习目标:“我能在感到沮丧或不堪重负时调节自己的反应”/目的:个人反思/student_generated:是/未指定等级区间
单点量规:自我调节反思
**学习目标:**我能在感到沮丧或不堪重负时调节自己的反应 **目的:**个人反思 **所选格式:**学生生成式——student_generated 为“yes”
学生生成的提示词
给学生的引导语: 想一想,当你感到沮丧或不堪重负时,调节自己的反应对你个人来说究竟是什么样子。写一个句子来描述它——不是一般情况下的样子,而是你自己做得好的时候是什么样子。
支架提示:
- “如果有人在观察你,他们能看到或听到什么,从而知道你正在调节自己的反应,而不只是把情绪压下去或突然爆发?”
- “想一想你曾经妥善应对沮丧或不堪重负的一次经历。你具体做了什么?那一刻和处理得不好的时候有什么不同?”
- “要让你说出‘我调节了自己的反应’,最重要的一个条件是什么?不只是保持安静,而是真正处理好了自己的情绪?”
给教师的说明: 学生生成的标准应当是个人化的,并且针对该学生。不要纠正、标准化,也不要用学习目标的语言替换它。学生写下“我可以先数到十并深呼吸,然后再说话”,这是一个有效且具体的个人标准——它体现了该学生的具体策略,以及对他们来说调节反应是什么样子。这种具体性正是其价值所在。对于年龄较小的学习者,或觉得任务有难度的学生,可以通过对话共同生成标准:大声读出支架提示,并帮助学生把回答组织成一个句子。
已知局限
-
要写好标准陈述,需要作出判断。 过于宽泛的标准(“作品体现了有质量的思考”)会产生“我认真思考了”这样的证据——毫无用处。过于狭窄的标准(“论证恰好包含三条证据”)则会变成检查清单。这项技能会生成一个标准和一项质量检查,但教育者仍应阅读两者,并在标准仍不合适时进行调整。
-
单点式量规是形成性工具,而非入门工具。 学生已经接触过学习目标,并对相关领域有一定经验时,单点式量规效果最好。一个从未写过论证的学生,无法有意义地对论证标准进行自我评估。应在教学之后使用这项技能,而不是之前。
-
学生生成版本的实用性取决于学生对自己的了解。 对于年龄较小的学习者(Band A–B,5–9 岁),独立生成标准在发展阶段上要求较高。通过对话共同生成会更合适——支架提示旨在支持这一过程,但教师不应指望学生在没有支持的情况下写出完整的句子。
-
这项技能不会生成成绩或总结性等级。 填写完成的量规各栏是用于交流、反思和规划后续步骤的证据。如果在单元或项目结束时需要作出总结性判断,请改用 coherent-rubric-logic-builder 或 criterion-referenced-rubric-generator。
-
三栏格式要求学生能同时关注两个方向——成长的证据和优势的证据。有些学生,尤其是有完美主义倾向或自我效能感较低的学生,会难以如实填写优势栏,或会把成长栏当作自我惩罚的工具。向已知面临自我效能挑战的学生发放此量规时,应明确说明:“右栏不是自夸,而是证据。你的任务是找到这些证据。”