bootloops-ai/bootloops-research
The BootLoops research skills: heavier machinery for specific tasks — proving with agents (adversary-first pipelines), pre-circulation referee audit, the literature review behind a novelty claim, bibliography verification, and the scientific-prose linter.
Deep literature review behind a research project's novelty claims — a protocol whose one job is to prevent shortcuts. It demands the papers be actually read (nothing from memory), makes thoroughness falsifiable (read ledgers, search receipts, gap statements), tracks citation chains in both directions, and lists the places to look. Use before any novelty claim ships, when writing or reviewing a related-work section, and when refereeing.
Multi-agent protocol for proving or refuting mathematical statements. Use when trying to prove a mathematical statement with agents, when a claimed lemma keeps dying under scrutiny, or when the true statement itself is still unknown (theory-finding regime).
Careful bibliography verification and entry-addition for papers — verify every author, title, journal, volume, pages, year, arXiv ID, and DOI against authoritative online sources (INSPIRE, arXiv, CrossRef/DOI, zbMATH, publisher) before anything enters or stays in a .bib file. Use whenever adding bib entries, auditing a bibliography, checking citations, fixing references, or when asked to "check the references". Encodes the report-then-fix pattern for full-file audits and the folklore-error traps (wrong years, truncated author lists, mislabeled keys).
Use before circulating any outward-facing scientific document — simulate the strongest standard objection from every audience it claims and verify the document answers each where that reader would look.