whitespectre

eval-clarity

Score assistant responses for clarity on a strict 1-5 scale, then return strict JSON only with score, rationale, and improvement suggestions. Use when the user asks to evaluate clarity, grade clarity, or critique clarity quality.

whitespectre 0 Updated 3mo ago

Resources

1
GitHub

Install

npx skillscat add whitespectre/ai-assistant-evals/eval-clarity

Install via the SkillsCat registry.

SKILL.md

Eval Clarity

Use this skill to evaluate how clear and understandable an assistant response is.

Inputs

Require:

  • The assistant response text to evaluate.

Internal Rubric (1–5)

5 = Structured, unambiguous, direct answer, minimal fluff
4 = Mostly clear, minor ambiguity or verbosity
3 = Understandable but lacks structure or precision
2 = Vague, missing key steps, hard to follow
1 = Confusing, contradictory, or unclear

Workflow

  1. Assess structure, precision, and readability.
  2. Score clarity on a 1-5 integer scale using the rubric only.
  3. Write concise rationale tied directly to rubric criteria.
  4. Produce actionable suggestions that improve clarity.

Output Contract

Return JSON only. Do not include markdown, backticks, prose, or extra keys.

Use exactly this schema:

{
"dimension": "clarity",
"score": 1,
"rationale": "...",
"improvement_suggestions": [
"..."
]
}

Hard Rules

  • dimension must always equal "clarity".
  • score must be an integer from 1 to 5.
  • rationale must be concise (max 3 sentences).
  • Do not include step-by-step reasoning.
  • improvement_suggestions must be a non-empty array of concrete edits.
  • Never output text outside the JSON object.