dozybot001

refined-idea-quality

Quality criteria for refined_idea output. Use before RefineIdea or when ValidateRefinedIdea returns low score. Focus on concrete, decomposable, and specific output.

dozybot001 7 Updated 3mo ago
GitHub

Install

npx skillscat add dozybot001/maars/refined-idea-quality

Install via the SkillsCat registry.

SKILL.md

Refined Idea Quality Criteria

Guidelines for producing a high-quality refined idea (Markdown) that is decomposable into 3–10 tasks. Quality matters.

Quality Rules

  1. Concrete over vague: "Compare BERT and GPT on code completion benchmarks" not "Research NLP models".
  2. Combine literature insights with novel angles: Avoid mere summarization; add synthesis, comparison, or gap.
  3. Must be decomposable: If you cannot list 3+ distinct sub-tasks, refine further.
  4. Gap must be specific: Name limitations (e.g. "prior work focuses on X, neglects Y") rather than generic "more research needed".

Validation Hints (ValidateRefinedIdea)

  • Score >= 4: Proceed to FinishIdea.
  • Score < 4: Call RefineIdea again with improvements — tighten scope, clarify questions, or sharpen gap.

Red Flags

  • Output contains "research" without scope
  • Questions are yes/no or too broad
  • Gap is generic ("limited studies", "need more work")
  • Method or approach is a single vague step