WorldwideTeaching

Build a criterion-referenced assessment rubric

Draft observable performance descriptors for a supplied assessment task, with a student version and notes for calibrating marking decisions.

Inspect source ↗
Work typeTeaching
SetupInstruction-led
ContextWorldwide

Is this the right workflow?

TASK GUIDE

A useful fit

Preparing a task-specific marking guide or making vague performance descriptors observable and easier for students and markers to discuss.

Boundaries and review decisions

A generated rubric is not a validated grading instrument. Educators decide criteria, weightings, fairness and accommodations. The publisher directs Manning competency programmes to a different companion skill; this listing covers general curriculum tasks.

Prepare a first draft

Fieldwork preparation plan adapted from the pinned publisher workflow. Read the complete pinned source before use.

The guidance below adds task-specific review limits and replaces silent defaults with explicit supplied context. It is a preparation plan, not a report of an executed skill.

Attribution and reuse

Adapted from Gareth Manning’s criterion-referenced-rubric-generator. Fieldwork changed the task summary, preparation steps, review limits and starter prompt. This listing’s adapted task guidance and prompt are licensed under CC BY-SA 4.0; this applies to those materials, not unrelated site content.

Licence terms · Copyright and permission notice

  1. Define the evidence to assess

    Supply the task, objective and student level. Agree the criteria and performance levels before drafting; keep unrelated presentation preferences out unless they are assessed outcomes.

  2. Describe observable differences

    Draft each criterion separately with concrete evidence for each level. Check that adjacent levels differ in demonstrated performance rather than adjectives alone.

  3. Calibrate before use

    Compare the draft against permitted samples and discuss ambiguous cases with markers. Review the student version and revise inconsistent descriptors before marking live work.

Fieldwork task guidance updated 2026-10-20. Publisher source inspection remains dated 2026-10-20; independent execution testing has not been performed.

01 / What you bring

REQUIRED INPUTS

Gather the material before running the prompt. Follow the publisher’s full checklist for the specific task.

INPUT 01

Input 01

Learning objective and the exact assessment task

INPUT 02

Input 02

Student year level, subject and applicable assessment framework

INPUT 03

Input 03

Approved criteria, number of performance levels and permitted calibration samples

Publisher excerpt: “Generate a criterion-referenced rubric with descriptive performance levels for a task or objective. Use for marking guides and…” Source ↗

02 / What you get

EXPECTED OUTPUT

An assessment rubric, criterion rationale, student-friendly version and marker-calibration notes for educator review.

FIRST-TRIAL CHECKLISTNO EXECUTION RESULT CLAIMED

Choose a controlled input

Learning objective and the exact assessment task; Student year level, subject and applicable assessment framework; Approved criteria, number of performance levels and permitted calibration samples

Check it before relying on it

Ask two educators to discuss the same permitted sample against the draft. Record disagreement and revise descriptors; do not invent calibration scores or reliability measures.

Use a small, permitted sample and compare the result with a reviewed baseline. This plan does not claim the publisher’s skill has been run or measured by Fieldwork.

03 / Set it up

Claude: follow the publisher’s Claude or Claude Code installation path. Preserve the complete skill folder and supporting references; configure required tools before trying the prompt.

  1. STEP 01

    Review the complete source

    Inspect the pinned source and licence notice before use.

  2. STEP 02

    Keep the full workflow

    For a compatible Agent Skills host, use the complete folder; for ChatGPT, supply the instruction file manually and use the starter prompt.

  3. STEP 03

    Configure your workspace

    Provide the named inputs and confirm the task-specific limits before drafting.

  4. STEP 04

    Try a reviewable first draft

    Review a small permitted sample before applying the workflow to live work.

Compatibility, tools & dependencies

Published as an Agent Skills instruction file for compatible Claude setups. ChatGPT use is a manual prompt/instruction adaptation, not a native installation or publisher-tested execution. No Fieldwork execution test was performed.

Supply the actual Australian curriculum or training framework and year level. Educator review is required; this is not a curriculum-compliance certification.

Named tools: See publisher notes. Actual access depends on your setup.

  • The complete pinned SKILL.md and publisher licence notice. No bundled scripts or mandatory external connector in this folder.
  • The supplied task evidence and a responsible human reviewer. Optional companion skills are separate, not installed by this listing.

Pinned source and complete folder inspected on 20 October 2026. No installation, script execution or independently measured result.

Public instructions do not imply free access to AI products, apps or APIs.

Publisher installation notes ↗ Platform setup guide →

04 / Your starter prompt

REPLACE THE BRACKETS
STARTER_PROMPT.TXT / REV 6bbbce4Download .txt ↓
Pinned publisher workflow: https://github.com/GarethManning/education-agent-skills/blob/6bbbce418f82e11044009c9f3b7373a354de5bd0/skills/curriculum-assessment/criterion-referenced-rubric-generator/SKILL.md

Task: Build a criterion-referenced assessment rubric.
Use the complete source instructions, subject to the task-specific review limits below.

Supply:
- Learning objective and the exact assessment task
- Student year level, subject and applicable assessment framework
- Approved criteria, number of performance levels and permitted calibration samples

First identify missing evidence and ask for it rather than assuming defaults. Use only authorised materials.

Preparation sequence:
1. Define the evidence to assess: Supply the task, objective and student level. Agree the criteria and performance levels before drafting; keep unrelated presentation preferences out unless they are assessed outcomes.
2. Describe observable differences: Draft each criterion separately with concrete evidence for each level. Check that adjacent levels differ in demonstrated performance rather than adjectives alone.
3. Calibrate before use: Compare the draft against permitted samples and discuss ambiguous cases with markers. Review the student version and revise inconsistent descriptors before marking live work.

Return: An assessment rubric, criterion rationale, student-friendly version and marker-calibration notes for educator review.
Review limits: A generated rubric is not a validated grading instrument. Educators decide criteria, weightings, fairness and accommodations. The publisher directs Manning competency programmes to a different companion skill; this listing covers general curriculum tasks.
First-trial check: Ask two educators to discuss the same permitted sample against the draft. Record disagreement and revise descriptors; do not invent calibration scores or reliability measures.

Separate supplied facts, calculations, assumptions and unresolved questions. Do not invent examples, measured results or access. Return the draft here for human review; do not send, publish, change records or make purchases. If the source or required evidence is unavailable, state the limit.

Attribution and reuse: Adapted from Gareth Manning’s criterion-referenced-rubric-generator. Fieldwork changed the task summary, preparation steps, review limits and starter prompt. This listing’s adapted task guidance and prompt are licensed under CC BY-SA 4.0; this applies to those materials, not unrelated site content.
Licence: https://creativecommons.org/licenses/by-sa/4.0/
Copyright and permission notice: https://findaiskills.com.au/source-licences/gareth-manning-cc-by-sa-4.0.txt

Fieldwork’s starter prompt. Keep the complete publisher workflow and replace the brackets before use.

05 / Before you rely on it

Review the draft before using it

Ask two educators to discuss the same permitted sample against the draft. Record disagreement and revise descriptors; do not invent calibration scores or reliability measures. A generated rubric is not a validated grading instrument. Educators decide criteria, weightings, fairness and accommodations. The publisher directs Manning competency programmes to a different companion skill; this listing covers general curriculum tasks.

  • Ask two educators to discuss the same permitted sample against the draft. Record disagreement and revise descriptors; do not invent calibration scores or reliability measures.
  • A generated rubric is not a validated grading instrument. Educators decide criteria, weightings, fairness and accommodations. The publisher directs Manning competency programmes to a different companion skill; this listing covers general curriculum tasks.

Read the related practical guide →

This is a source-inspected starting point. Fieldwork has not run-tested this skill or measured time savings.