A Claude Code skill that enables any LLM to write in the style of László Krasznahorkai's Satantango — with dramatically higher information density and reading efficiency.
Satantango is the masterwork of László Krasznahorkai, winner of the 2025 Nobel Prize in Literature.
mkdir -p ~/.claude/skills/satantango
curl -o ~/.claude/skills/satantango/SKILL.md \
https://raw.githubusercontent.com/28222243tfwfsaf/satantango-skill/main/satantango.mdMost writing advice tells you to use short sentences. This is wrong — for information delivery. Short sentences waste words rebuilding context at every period. Each new sentence forces the reader to re-establish who is speaking, where they are, and what frame they're in.
Krasznahorkai's long-sentence technique solves this:
- One frame, many data points. Build the scaffolding once. Accumulate. No redundant subject-verb reconstructions.
- Parataxis compresses. Juxtaposed clauses carry more information per word than separated sentences.
- No micro-reorientations. The reader stays inside the consciousness. Every period is a wall — the reader must climb over it and re-locate.
- Simultaneity preserved. Physical detail, psychological state, and philosophical claim coexist in syntax, just as they do in reality.
- Breath is the natural unit of thought. We don't think in periods. We think in clauses, turns, self-revisions. The prose reproduces a mind working.
The result: higher information density per sentence, faster reading comprehension, less cognitive overhead.
A pure text-based writing instruction set — not an app, not an API, not a library. Based on deep bilingual analysis of the English translation (George Szirtes) and the Chinese translation (余泽民, Yilin Press, 2017). Contains no copyrighted text — this is original writing craft analysis.
MIT — see LICENSE.
Analysis based on:
- Sátántangó by Krasznahorkai László (1985)
- English translation Satantango by George Szirtes (New Directions, 2012)
- Chinese translation《撒旦探戈》by 余泽民 (译林出版社, 2017)