ai-scientist-skills

Generate a complete LaTeX research paper from experiment results — including citation gathering, figure descriptions, and iterative refinement.

Run the complete 4-stage Best-First Tree Search experiment pipeline — from initial implementation through ablation studies, with multi-agent parallel exploration.

Execute a single Best-First Tree Search iteration — select parent node, generate experiment code, execute it, parse metrics, analyze plots, and update the journal.

Run the complete AI Scientist pipeline — from research ideation through experiment execution, paper writing, and peer review. Orchestrates all sub-skills.

Interactively guide the user to create a workshop/topic description file for the AI Scientist pipeline. Asks structured questions and generates a ready-to-use .md file.

Generate novel research ideas with literature search and novelty checking. Produces structured JSON ideas matching the AI Scientist schema.

Search academic literature using Semantic Scholar API or WebSearch. Returns papers with titles, abstracts, citation counts, and optionally bibtex entries.

Aggregate and generate publication-quality figures from experiment results across all BFTS stages.

Perform a structured peer review of a research paper — text analysis, figure quality assessment, and NeurIPS-format scoring.