📚 Education
Assignment Rubric and Grading Comment Bank
Build a standards-aligned rubric and a comment bank for an assignment, including level descriptors and rewrite-ready feedback.
0Reviews
Prompt
Act as a classroom teacher writing a rubric students can actually use while drafting, plus comments a tired grader can paste. Align to the stated standard. Do not invent state-standard codes. Inputs: - Assignment: [Name, task] - Grade / course: [Grade] - Standard or outcome: [Text, not a guessed code] - Points / weight: [Points] - What excellence looks like: [If you have an exemplar note] - Common failures: [List] - Comment voice: [Warm / Direct] - Must-include: [Citations, showing work, lab safety, etc.] Generate: 1. Rubric table: 4 criteria x 4 levels (Exceeds / Meets / Approaching / Beginning). Observable verbs. Points that add to the given total. One criterion is the Must-include. 2. Student-facing checklist: 8 yes/no items they can run before turning in. 3. Comment bank: for each criterion, 3 comments (exceeds, meets, not yet) plus a rewrite stem the student can use. No sarcasm. 4. Whole-paper comments: 4 short paragraphs (strong, mixed, incomplete, missing Must-include). 5. Grading notes for the teacher: 5 edge cases (late, extra credit, group, AI-sounding, missing name). Policy only if Inputs gave one; else flag OPEN. 6. 10-minute grade path: what to read first so scores stay consistent. Constraints: - No made-up CCSS or NGSS numbers. - Criteria cannot be "effort" or "creativity" unless the assignment grades those with artifacts. - Comments tell the student what to change next, not who they are.
Instructions
Replace every [bracket] with your details before running. Works on ChatGPT, Claude, and Gemini.
Examples
Example Input
Assignment: 2-page lab report on the vacuum vs plastic bottle heat-loss lab Grade / course: 8th grade physical science Standard or outcome: Students explain a controlled comparison and use evidence from their own data table to support a claim about heat flow. Points / weight: 24 points What excellence looks like: claim matches the table; they name the one variable they changed; they admit we did not test 24 hours Common failures: copied procedure without data; claim that the bottle creates cold; no units; group name only, no individual conclusion Comment voice: Direct Must-include: data table with units, and a one-sentence limit (what we did not test)
Example Output
1. Rubric (24 pts) Criteria (6 each): Claim, Evidence (data table + units) [Must-include], Reasoning, Limit statement [Must-include split: table 4 + limit 2 if you need, but keep 6/6/6/6]. Claim 6/4/2/0: Matches table (vacuum lost less heat) / mostly / says bottle creates cold / missing. Evidence 6/4/2/0: Table with C, minutes, both bottles / numbers no units / copied procedure no table / none. Reasoning 6/4/2/0: Ties vacuum = poor conduction to the 4 C vs 19 C / names vacuum without the numbers / magic cold / blank. Limit 6/4/2/0: Names not-24h and not-copper / one limit / fake extra tests / none. 2. Checklist I stated a claim that matches my table. I did not say the bottle creates cold. Table has both bottles. Units on temperature and time. Start temp 80 C if that was ours. One thing we did not test. My conclusion is mine, not only the group chat. Procedure is short, data is long. 3. Comments (direct) Claim exceeds: Your claim matches the 4 C vs 19 C drop. Claim not yet: "Creates cold" is not in this lab. Rewrite: The vacuum bottle lost less heat in 30 minutes. Evidence meets: Table has units. Add the start temperature if it is missing. Evidence not yet: I cannot find a table. Paste the two bottles, minutes, and C. Reasoning not yet: Name the vacuum as a poor conductor, then point at your two numbers. Limit exceeds: You said we did not test 24 hours. Keep that sentence. Limit not yet: Add one line: We did not test 24 hours or copper lining. 4. Whole-paper Strong: Claim, table, reasoning, limit. Ready. Mixed: Table is there, claim overreaches. Fix the cold sentence. Incomplete: Procedure only. Missing must-include: No table or no limit sentence. Grade cannot exceed Approaching on those criteria. 5. Teacher edges Late: OPEN, no policy in Inputs. Extra credit: none unless you add it. Group: individual conclusion required. AI-sounding: compare to their table; if no 4 and 19, send back. Missing name: do not enter until named. 6. 10-minute path Find the table first. Then the cold-creates sentence. Then the limit. Then claim/reasoning. Scores stay on those four.