Assessment / Writing
The sentence.
The skill. The next step.
The student’s text is assessed against the assigned task and its rubric. Feedback points back to the writing, so a teacher can inspect what the engine has proposed.
What is assessed
Meaning and control.
The engine considers how the writing flows, the words chosen, grammatical and mechanical accuracy, sentence structure and fulfilment of the task. It retains original text alongside suggested corrections and skill tags.
Standard assignments can include specific teacher requirements. IELTS writing follows a separate prompt and rubric path for the selected task.
The Standard rubric
The report’s own labels.
| Criterion | What the reader should look for |
|---|---|
| Flow | Logical progression and connection between ideas. |
| Lexical | Vocabulary choice, precision and range. |
| Accuracy | Control of grammar, spelling and mechanics. |
| Structure | Sentence construction and syntactic variety. |
| Mission | How the writing fulfils the assigned task. |
Standard 50 uses criteria displayed out of 10; Standard 100 displays them out of 20. The total is assembled from the underlying values before display rounding. Off-topic work can be capped even when the language is polished.
templates/student_processing/pdf_components/_dashboard_report.htmlmodules/student_processing/services/pdf_service.pymodules/common/scoring_math.pyIELTS writing
Use the task’s rubric.
| Code | Criterion |
|---|---|
| TR / TA | Task Response for Task 2; Task Achievement for Task 1. |
| CC | Coherence and Cohesion. |
| LR | Lexical Resource. |
| GRA | Grammatical Range and Accuracy. |
These are band-based assessments on the IELTS path. A converted label on a Standard report is an estimated equivalency, not an official IELTS result.
scripts/seeds/scales/data_v2.py · IELTS Task 1 / Task 2docs/reference/IELTS_TASK1_GLOBAL_POLICY.mdInside the report
Enough context to review.
- Rubric scores and an overall profile.
- Language strengths and a Precision Audit of original text, correction and diagnosis.
- A learning path that identifies a pattern and the next practice goal.
- English and translated feedback where that report configuration supplies it.
The real artefact
Read the sample.
These are the supplied writing report pages, converted to WebP without altering their contents. Individual scores and equivalencies belong to this stored example; they are not validation results.
next-web/public/sample-report/writing-1.jpg … writing-5.jpgdocs/architecture/domain_writing_unified.mddocs/architecture/04_PROMPT_ENGINEERING_AND_TAXONOMY.md