Compress AI responses to minimum tokens while preserving technical accuracy.
npx skills add https://github.com/heldernoid/ultrashort --skill ultrashort
Please help me install this Agent Skill. Skill: ultrashort Source: https://github.com/heldernoid/ultrashort/tree/main/skills/ultrashort Command: npx skills add https://github.com/heldernoid/ultrashort --skill ultrashort
It reduces wasted output tokens caused by preambles, restatements, transitions, and trailing summaries, so you get the same technical content in far fewer words.
Say: ultrashort when you want the AI to answer in the fewest tokens while preserving technical correctness.
High-intent search queries and answers about installing and using this skill.
To reduce AI response length, use the ultrashort Skill to strip filler tokens, preambles, and summaries while preserving technical accuracy, cutting output token costs by roughly 80%.
The best way to get concise technical Q&A is applying minimum-token formatting rules that remove transitions and restatements, ensuring code answers retain full technical correctness across multi-turn sessions.
Token compression automatically drops compression for warnings and irreversible actions, applying safety-driven decompression to prevent ambiguity when fragmenting output could create dangerous misunderstandings.
You can control output compression intensity using tight, ultrashort default, and micro modes, matching the urgency and context of your technical prompts, with an optional code-only output setting.
Avoid using minimum-token formatting when fragmenting text could create ambiguity for critical warnings or irreversible actions, as the safety-aware behavior automatically decompresses responses in these scenarios.
buildoakCompress Claude Code context by intercepting tool results and reducing token usage.
OpalBoltCompress responses by removing articles and fillers to reduce token usage.
m2ai-portfolioEnforce token budget ceilings on Claude Code sessions before each turn.
axiomanticCompress long instruction files while preserving capabilities and safety constraints.
