Generate ranked empirical research ideas from the EmpiricalWiki knowledge graph.
npx skills add https://github.com/Lambenthan/empiricalwiki --skill ideate-lambenthan
Please help me install this Agent Skill. Skill: ideate Source: https://github.com/Lambenthan/empiricalwiki/tree/main/.claude/skills/ideate Command: npx skills add https://github.com/Lambenthan/empiricalwiki --skill ideate-lambenthan
EmpiricalWiki users need a reliable way to convert scattered papers, existing claims, and knowledge gaps into concrete, testable research ideas without repeatedly reinventing the same directions.
Run the ideate skill with your research direction to generate, validate, and write the top research ideas into the wiki.
High-intent search queries and answers about installing and using this skill.
You can generate testable empirical research ideas by scanning a knowledge graph and external literature, applying dual-model brainstorming and feasibility filtering to target specific open gaps. This pipeline outputs 3–6 ranked ideas ready for experiment design.
Validating research novelty and feasibility involves running deep review checks against prior claims and a banlist of failed ideas. This screening process ensures new candidates are original, achievable within 3–6 months, and explicitly aligned with open knowledge gaps.
You can prevent duplicate research ideation by loading a wiki-grounded banlist of prior failed ideas into the pipeline. This anti-repetition memory records specific failure reasons and blocks the generation of previously explored directions.
Writing validated research ideas into a wiki knowledge graph involves creating dedicated idea pages and adding relational edges connecting the ideas to relevant gaps, claims, and papers. This keeps the graph queryable for future ideation cycles.
Automated research idea generation pipelines require multi-phase execution and may allow optional validation skipping, which risks lower quality outputs. Additionally, they depend heavily on the existing wiki graph structure and may not fully replace manual expert domain review.
LambenthanRetrieves wiki pages and synthesizes evidence-grounded answers with citations.
hyperbolic-cGenerate structured, testable hypotheses with mechanistic explanations and falsifiable predictions from observations and literature syntheses.
LambenthanAutomate daily arXiv paper discovery, relevance scoring, and controlled ingest into a research wiki.
bacocoGenerate structured discovery briefs and research dossiers from vague ideas.
