Analyzes research ideas and generates theory drafts, experiment plans, and critical reviews.
npx skills add https://github.com/megumi-ben/work17 --skill check-idea-megumi-ben
Please help me install this Agent Skill. Skill: check-idea Source: https://github.com/megumi-ben/work17/tree/main/.codex/skills/check-idea Command: npx skills add https://github.com/megumi-ben/work17 --skill check-idea-megumi-ben
theory.md, with KaTeX-compatible formatting. - Experiment Planning & Critical Review: Designs an experiment framework (baselines, ablations, robustness) and lists the weakest assumptions and likely reviewer objections in check_idea.md. - Use Case: A PhD student with a rough idea and several related papers asks the assistant to check the idea; the Skill writes a staged analysis into check_idea.md and theory.md suitable for NeurIPS submission planning. ## Quick Start Use the check-idea skill to analyze my research idea based on the attached reference papers and write the results to check_idea.md and theory.md.High-intent search queries and answers about installing and using this skill.
Provide your reference files and ask for an idea check. The Skill extracts the core contribution, contrasts it with common approaches, and explicitly assesses novelty and whether the idea could target a NeurIPS oral.
The Skill's stage 2 writes a full theoretical analysis to theory.md, including problem setup, notation, formal objectives, derivations, and propositions or proofs, while flagging which parts are rigorous versus heuristic.
It produces two markdown files: check_idea.md containing the idea summary, experiment plan, and critical review, and theory.md containing the complete theoretical derivation. You can override these default filenames.
Yes. The default outputs are check_idea.md and theory.md, but if you specify different output paths, the Skill writes the staged content to your chosen files instead.
The analysis depends entirely on the reference files and idea description you provide, so missing related work can weaken novelty judgments. It also cannot run experiments, so empirical claims remain planned hypotheses rather than verified results.
JaimeParkerSearch the arXiv repository for academic papers using the arxiv Python package.
lightrain-aGenerate a structured paper outline from review conclusions and experiment results.
mnazaalEnforces verification of academic papers, citations, and literature claims before citing.
hxk622Generates novel research hypotheses and interdisciplinary connections through conversational brainstorming with web search and code execution.