reading-level

Estimate a passage's grade band and Lexile. Do NOT use for whole documents or student data.

SKILL.md

GitHub

reading-level

Estimates reading level of a text passage and returns a grade-band label, approximate Lexile range, and a brief rationale. Zero-token path when the readability library is installed (Flesch-Kincaid); model-based estimate otherwise. Used by lesson-planner, assessment-item, and differentiate atoms to verify appropriateness before passing content downstream.

Input

{
  "text": "The water cycle describes how water moves...",
  "target_grade": "5",
  "context": "science passage for a quiz"
}

target_grade: the intended audience grade. context (optional): helps calibrate the estimate.

Output

{
  "tool": "reading-level",
  "estimated_grade_band": "4-5",
  "lexile_estimate": "700-750L",
  "match_for_target": true,
  "rationale": "Sentence length averages 14 words; academic vocabulary present but not dense; Flesch-Kincaid grade ~4.8.",
  "method": "flesch_kincaid | model_estimate",
  "recommendation": "Appropriate for grade 5. For ELL/below-grade readers, simplify vocabulary in sentences 2 and 4.",
  "human_review_required": true
}

match_for_target: true if the estimated band includes the target grade. method is flesch_kincaid when the readability or textstat library is installed, model_estimate otherwise (less precise).

Do NOT use this atom for

  • Full document readability audits (call this atom once per passage)
  • Student performance data or reading assessment scores
  • Assigning a Lexile as a guaranteed authoritative measure — it's an estimate

Pipeline note

Follows references/method.md at the Analysis step (reading-level estimation). Output conforms to references/metadata-schema.md. Standards and Differentiation steps are skipped. This atom is a quality pre-check only. human_review_required: true — Lexile estimates are directional, not certified.