reading-use-of-english

Reading & Use of English

Objective drills: generate → answer → score → explain → log. Paths are relative to ${CLAUDE_PLUGIN_ROOT} (if unset, resolve relative to the plugin root — the directory two levels above this file, i.e. ../../ from here).

When to use

The user asks to drill any reading or use-of-english task type, or asks why a reading answer is what it is.

Steps

1. Identify exam + level + task type (ask once if unclear; /daily-drill picks the weakest type from the progress log instead). Load data/exam-formats/<exam-id>.md for the exact shape (items per task, options per question) and data/cefr/reading-descriptors.md + data/cefr/reading-calibration-anchors.md to calibrate text difficulty. Seed shapes: data/item-bank/seed/reading-use-of-english-items.md. Check the requested task type is actually in the chosen exam (per its format file): key word transformation and word formation exist only in B2–C2, cross-text matching only in C1, and so on. If the user asks for a task their exam does not contain, say so plainly and offer the nearest valid task or name the exam(s) that do include it — never generate an off-format item. True/False/Not Given vs Yes/No/Not Given: T/F/NG statements are about FACTUAL information in the text; Y/N/NG statements are about the WRITER'S views/claims (so the passage must carry opinions). Use the correct labels for each.

2. Generate an ORIGINAL passage and items matching the format exactly: right number of items, right option count, plausible distractors, one defensibly correct answer each. Calibrate lexis and syntax to the CEFR level. Give a realistic time budget (e.g. ~1.3 min per use-of-english item; proportional share of the section time for reading sets).

   Match the passage to authentic length — this is as important as the item count. Aim for the upper end of the target range and count your words before presenting: left to instinct these passages come out ~30–40% too short. Use the exam's figure from data/exam-formats/<exam-id>.md ("Passage lengths"); as a fallback:

   • Gapped text: B2 ~500–600, C1 ~550–780, C2 ~700–800 words of base text.
   • Long-text multiple choice / multiple matching: B2 ~500–700, C1 ~700–800, C2 ~700–800 words of base text. IELTS Academic passages ~700–900 words each; IELTS General Training sections vary — see ielts-general.md. A too-short passage is the most common failure — a C2 Part 6 gapped text must be ~750 words, not ~450.
   • TOEFL Complete the Words (~70 words, exactly 10 gaps): show the first 3–5 letters of each gapped word, then one underscore per missing letter, counted exactly — flow___ = flow + 3 = "flowers". Count the answer's letters and subtract the stem for every gap; an eyeballed count gives the learner a different puzzle from the key. Leave the opening sentence intact. Worked example: data/item-bank/seed/reading-use-of-english-items.md.
   • Cloze / word formation: ~150–220 words. TOEFL academic passage ~200; TOEFL daily-life texts 15–150.
   • Gapped text always has ONE more option than gaps (B2/C1/C2) or three more (B1 Part 4) — the surplus fits no gap and is a deliberate distractor. Removed paragraphs run ~50–80 words each.