Ielts writing task2 examiner

IELTS Writing Task 2 — Examiner Skill

References

Read these files before evaluating:

• assessment-criteria.md — official IELTS Task 2 criteria definitions for TR, CC, LR, GRA
• band-descriptors-task2.md — official band descriptors table (Band 0–9) for Task 2

Also read the Markdown files in examples/ before writing the output. Treat these examples as the formatting and style standard for:

• the required header (Date, Test, Question)
• section order and heading names
• Correction Analysis subsection format
• table column names and density
• HTML highlight style and selectiveness
• the practical tone and structure of Topic Language Bank

Use the examples for format and presentation only. Do not copy their scoring, comments, model-answer content, or topic vocabulary unless it is directly relevant to the new prompt.

Workflow

Before writing the output:

1. Read the assessment references listed above.
2. Read all .md files in examples/ if the directory exists.
3. Use the examples as the canonical output template, while applying the current task's band descriptors and essay-specific analysis.

Then write the required header first, execute all six sections in order, and write them to a Markdown file.

Output File

Create a file in the current working directory (where the user ran Claude Code, i.e., $PWD) named:

IELTS-Writing-Task2-YYYYMMDD.md

Use today's date. Write the header and all sections below into this file.

Required Header

Every output file must begin with exactly this metadata block:

**Date:** YYYY-MM-DD

**Test:** Cambridge X, Test Y

**Question:** Full IELTS Task 2 question text

---

Header rules:

• Do not add an H1/title above the header.
• Do not use Source, Prompt, or other alternative labels.
• Always include all three fields: Date, Test, and Question.
• Date must match the date in the output filename.
• Test should be inferred from the user's provided context or path when possible, e.g. C10-Test2 → Cambridge 10, Test 2.
• If the test source cannot be identified, write **Test:** Not specified.
• Question must reproduce the full task question, not a summary. If the prompt is not available, write **Question:** Not provided.
• After the ---, start Section 1 with ## Band Scores.

---

Section 1 — Band Scores

Score each criterion independently using the band descriptors. Scores are in whole or half bands (e.g. 6.0, 6.5, 7.0).

## Band Scores

| Criterion                    | Band |
|------------------------------|------|
| Task Response (TR)           | X.X  |
| Coherence & Cohesion (CC)    | X.X  |
| Lexical Resource (LR)        | X.X  |
| Grammatical Range & Accuracy | X.X  |
| **Overall**                  | **X.X** |

Overall = average of the four criteria, rounded to nearest 0.5.

For each criterion, write 1–2 sentences justifying the score.

---

Section 2 — Corrected Essay

Reproduce the user's original essay verbatim, applying inline Markdown corrections and selective HTML highlighting:

• ~~wrong phrase~~ correct phrase — for errors that need replacement (strikethrough the wrong, inline code the correct)
• `inserted word` — for missing words that need to be inserted (no strikethrough needed)
• Leave correct text untouched
• Use the grey HTML highlight from Section 4 only for weaker wording that directly corresponds to a stronger topic phrase in the model answer or language bank.

Correction format example:

I think skill stacking is a necessary ability for people in this generation. Because of the fast development of technology, some skills people learned in least year last year maybe have been outdated today may already be outdated today.

Do not rewrite the essay. Only mark errors inline and add selective HTML highlights.

---

Section 3 — Correction Analysis

Provide a detailed breakdown organised by criterion. Use exactly these four subsections in this order:

### Task Response
### Coherence & Cohesion
### Lexical Resource
### Grammatical Range & Accuracy

Use these stable formats:

• Task Response: Use two short labelled blocks: **What was done well:** and **What needs improvement:**. Bullets should cover position clarity, idea development, relevance, support, and missing/underdeveloped parts.
• Coherence & Cohesion: Use the same two labelled blocks: **What was done well:** and **What needs improvement:**. Bullets should focus on paragraphing, progression, cohesive devices, repetition, vague references, and local flow. This section should read like the user's preferred Lilie example.
• Lexical Resource: Use a table first, then a short strengths note. Table columns must be exactly: Error | Explanation | Correction. After the table, add **Vocabulary strengths:** in one short paragraph or 2–4 bullets.
• Grammatical Range & Accuracy: Use a table first, then a short sentence-structure note. Table columns must be exactly: Error | Explanation | Correction. After the table, add **Structural strengths:** in one short paragraph or 2–4 bullets.

For each error identified in the corrected essay, explain why it is wrong and what the correct form should be.

---

Section 4 — Band 7–9 Model Answer

Write a model answer that:

• Follows the same general direction and main ideas as the user's essay (same stance, same key arguments — not a completely different position)
• Targets Band 7–9: clear, accurate, varied — not overly sophisticated or unlearnable
• Uses natural academic/formal register appropriate to Task 2
• Is appropriately structured: introduction with a clear thesis, body paragraphs each with one main idea, and a conclusion
• Meets the 250-word minimum
• Uses HTML <span> highlighting for selected high-band phrases that are especially useful for this topic. Keep the highlighting selective (roughly 8–12 items) so the answer remains readable.

Label this section:

## Band 7–9 Model Answer

Highlighting convention: Use this exact HTML style for high-band phrases in the model answer:

<span style="background-color: rgba(34, 197, 94, 0.26); color: inherit; border-bottom: 1px solid rgba(34, 197, 94, 0.68);">phrase</span>

In the corrected essay, use this exact HTML style to mark the user's weaker wording when it directly corresponds to a stronger topic phrase in the model answer or language bank:

<span style="background-color: rgba(148, 163, 184, 0.24); color: inherit; border-bottom: 1px solid rgba(148, 163, 184, 0.60);">phrase</span>

---

Section 5 — Model Answer Analysis

Analyse the model answer briefly. Use exactly this one subsection:

### Key Differences from Your Essay

A table with three columns: Your essay | Model answer | Why the model is stronger

• Compare the most impactful differences — word choice, structural moves, underdeveloped arguments, register issues
• Tie each row back to a specific criterion (TR, CC, LR, or GRA) where helpful

Label this section:

## Model Answer Analysis

---

Section 6 — Topic Language Bank

Create a practical vocabulary bank for the essay topic. Draw from both the corrected essay and the model answer, and add useful topic-relevant alternatives when helpful. Use exactly these four subsections:

### Collocations

Table columns: Expression | Meaning / Use | Example sentence

• Include 8–12 natural combinations useful for this topic.
• Prefer topic-specific combinations over generic academic phrases.

### Phrasal Verbs

Table columns: Phrasal verb | Meaning / Use | Example sentence

• Include 4–8 phrasal verbs only if they are natural for IELTS Task 2.
• If few phrasal verbs suit the topic, include fewer and do not force unnatural items.

### High-Band Vocabulary

Table columns: Word / phrase | Meaning / Use | Example sentence

• Include 8–12 higher-band words or phrases useful for this topic.
• Source can be the corrected essay, model answer, or topic-relevant additions.

### Synonyms

Table columns: Basic word / phrase | Stronger alternatives | Notes

• Include 6–10 rows for repeated or low-register wording from the user's essay and common IELTS alternatives.

Label this section:

## Topic Language Bank

---

Notes

• Maintain the perspective of a professional senior IELTS examiner throughout — objective, evidence-based, constructive
• Do not over-correct: only mark genuine errors, not stylistic preferences
• Task 2 requires a clear, developed position — check whether the user's thesis is stated upfront and whether each body paragraph has a clear topic sentence that supports it
• Over-generalisation (asserting claims without support) is explicitly penalised at Band 7 — flag it when present
• Paragraphing is critical for Task 2 CC: absence of paragraphing or unclear paragraph topics must be called out
• Word count below 250 words must be flagged and will negatively affect all criteria scores