
from wjgoarxiv
Autonomous research and experimentation loop for iterative optimization of code, prompts, and configs.
Inspired by Karpathy's autoresearch, this skill transforms an LLM into an autonomous research agent. It generalizes the iterative process of ML training to any domain that has a measurable metric and a defined search space.
Core Workflow:
research.md.Key Subcommands:
/autoresearch:plan: Setup wizard to create research.md./autoresearch:debug & :fix: Scientific bug hunting and iterative error correction./autoresearch:predict: Multi-persona deliberation./autoresearch:security: Iterative STRIDE/OWASP audits./autoresearch:ship: Universal shipping workflows.This skill has not been reviewed by our automated audit pipeline yet.