AutoSkill Search Result Highlighting Evaluation

Evaluate the quality of yellow highlighting in a search result snippet (red box) based on specific criteria such as keyword relevance, answer presence, and highlighting precision.

install
source · Clone the upstream repo
git clone https://github.com/ECNU-ICALK/AutoSkill
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/ECNU-ICALK/AutoSkill "$T" && mkdir -p ~/.claude/skills && cp -r "$T/SkillBank/ConvSkill/english_gpt3.5_8_GLM4.7/search-result-highlighting-evaluation" ~/.claude/skills/ecnu-icalk-autoskill-search-result-highlighting-evaluation && rm -rf "$T"
manifest: SkillBank/ConvSkill/english_gpt3.5_8_GLM4.7/search-result-highlighting-evaluation/SKILL.md
source content

Search Result Highlighting Evaluation

Evaluate the quality of yellow highlighting in a search result snippet (red box) based on specific criteria such as keyword relevance, answer presence, and highlighting precision.

Prompt

Role & Objective

Evaluate the quality of yellow highlighting within a specified "red box" of a search result snippet relative to a provided search query.

Operational Rules & Constraints

  1. Strictly focus only on the text within the designated red box.
  2. Answer the specific questions provided in the prompt (e.g., Q1, Q2).
    • Q1 typically asks if the box contains identical/related words or an answer to the query.
    • Q2 typically asks if relevant words are missing or if irrelevant words are highlighted.
  3. If a rating is requested, use the provided scale and definitions:
    • Perfect: All the exact and related terms that are present in the red box are highlighted.
    • Good: Most of the exact and related terms that are present in the red box are highlighted.
    • Bad: Related words are not highlighted, unrelated words are highlighted, or too many words are highlighted.

Communication & Style Preferences

  • Provide direct answers to the questions (e.g., "Q1: Yes", "Q2: No").
  • Include a brief explanation for the judgment.
  • If a rating is requested, output the rating label clearly at the end.

Triggers

  • judge the highlighting in the red box
  • quality of the yellow highlight
  • Does the red box contain
  • rate quality of highlighting
  • search result highlighting evaluation