Idea Generation

This is a structured workflow for generating research ideas with built-in reality checks. It walks you through generating 3-5 ideas, iteratively refining them for up to 5 rounds each, then running novelty checks against Semantic Scholar to see if your brilliant insight already exists in the literature. Each idea gets scored on interestingness, feasibility, and novelty, with the explicit goal of being conference-paper worthy. The included Python script automates the literature search part, which is honestly the most tedious bit of validating whether an idea is worth pursuing. It's opinionated about keeping ideas feasible (no massive datasets or compute) and not overfitting to specific benchmarks. Best used early in a research project when you're still figuring out what to work on.

skills.sh installs: 1.9K
GitHub stars: 459
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add lingzhi227/agent-research-skills --skill idea-generation --agent claude-code

Installs into .claude/skills of the current project.

Files

Idea Generation

Generate and refine novel research ideas with literature-backed novelty assessment.

Input

$0 — Research area, task description, or existing codebase context
$1 — Optional: additional context (e.g., "for NeurIPS", constraints)

Scripts

Novelty check against Semantic Scholar

python ~/.claude/skills/idea-generation/scripts/novelty_check.py \
  --idea "Adaptive attention head pruning via gradient-guided importance" \
  --max-rounds 5

Performs iterative literature search to assess if an idea is novel.