Wiki Research Loop

Runs an autonomous breadth-first research loop that grows a wiki by fetching from web, arXiv, and GitHub, deduping claims against existing pages, and queuing follow-up topics until you hit a budget cap, depth limit, or convergence threshold. Each iteration costs you tokens, so the built-in budget enforcement (defaults to 50 cents) and kill-switch file are your safety nets. The convergence detector halts when three consecutive pages add less than 5% novel content, which works better than you'd expect for focused technical domains. Use it when you want a knowledge base that builds itself overnight or need to map out a research area without manually chasing every citation. Privacy guard blocks external fetchers if your wiki is marked private.