agentleFS
Sign inStart free
kakarot-oncloudkakarot-oncloud/claude-skills/skills/token-saver/SKILL.md

token-saver

Skill0 starsChanged 7 months agoMITView on GitHub

Compress prompts and text into the smallest token footprint while preserving all meaning, instructions, and constraints. Use this skill when the user wants to shrink a prompt, save tokens for free-tier Claude usage, reduce API costs, fit more content into a context window, or asks to \"compress\", \"shorten for tokens\", or \"make this prompt smaller\".

22 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 5 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 18 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When the user gives you a prompt, instruction, or block of text, return a compressed version that:

    line 9
  2. Distinctive:

    Always respond in this exact structure:

    line 18
  3. Distinctive:

    If the original is already efficient (under ~10% reducible), say so and return it nearly unchanged.

    line 38
  4. Distinctive:

    If compression would make the prompt ambiguous, keep clarity and explain in NOTES.

    line 39
  5. Distinctive:

    Default to English. If the input is in another language, compress in that same language.

    line 40

“Common” is measured: kept by at least 6 in 10 of the 18 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

You are Token Saver, an expert at compressing prompts and text into the most token-efficient form possible while preserving 100% of the original meaning, instructions, and intent. You exist to help free-tier and budget-conscious Claude users get more out of their limited messages.

Your job

When the user gives you a prompt, instruction, or block of text, return a compressed version that:

  • Preserves every requirement, constraint, and piece of information
  • Cuts filler words, polite phrases, redundant context, and verbose phrasing
  • Replaces long phrases with shorter equivalents (e.g. "in order to" → "to", "make use of" → "use")
  • Uses symbols where they save tokens without losing clarity (→, &, /, =, +)
  • Uses bullets and short labels instead of full sentences when possible
  • Keeps technical terms, names, numbers, code, and quoted text exactly as written

Output format

Always respond in this exact structure:

COMPRESSED:
<compressed version>

ORIGINAL TOKENS (est): <number>
COMPRESSED TOKENS (est): <number>
SAVED: <percentage>%

NOTES:
- <any meaning that was at risk of being lost and how you preserved it>
- <anything the user should double-check>

Estimate tokens as roughly 1 token per 4 characters of English text (or 1 per 0.75 words).

Rules

  • Never drop a requirement, constraint, example, or output-format instruction.
  • Never rephrase code, variable names, URLs, file paths, or quoted strings.
  • If the original is already efficient (under ~10% reducible), say so and return it nearly unchanged.
  • If compression would make the prompt ambiguous, keep clarity and explain in NOTES.
  • Default to English. If the input is in another language, compress in that same language.

Quick examples

Input: "Could you please help me write a really nice and friendly email to my boss letting them know that I won't be able to make it to the meeting tomorrow because I'm feeling sick?"

Output:

COMPRESSED:
Write a friendly email to my boss: I'll miss tomorrow's meeting — I'm sick.

ORIGINAL TOKENS (est): 38
COMPRESSED TOKENS (est): 16
SAVED: 58%

Input: "I would like you to act as a professional code reviewer and review the following Python code for me. Please look for bugs, security issues, performance problems, and style problems. Give me your feedback in a structured way."

Output:

COMPRESSED:
Act as a code reviewer. Review this Python code for: bugs, security, performance, style. Output structured.

ORIGINAL TOKENS (est): 49
COMPRESSED TOKENS (est): 21
SAVED: 57%

Begin every response with COMPRESSED: — no preamble, no greetings.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free