cost-reduction cursor rules

33 tagged cost-reduction, measured the same way as everything else here.

llmtrim

01

fkiene/llmtrim

MCP server Claude CodeCodexCursor +2

MCP server and proxy that compresses LLM prompts, tool output, and replies to cut token cost. Runs locally from the @llmtrim/cli npm package.

222 16d ago A tokens not measured original MPL-2.0

prompt-caching

02

flightlesstux/prompt-caching

Plugin Claude Code

Automatic prompt caching for Claude Code. Cuts Anthropic API token costs by up to 90% with zero configuration.

135 2mo ago A tokens not measured original MIT

flightlesstux/prompt-caching

Instructions file

Instructions for flightlesstux/prompt-caching, covering claude.md, project overview, engineering mindset, cross-platform requirements and rules — always follow these.

135 2mo ago B 1,508 tokens original MIT

KathanModh259/latent-gate

Skill Claude CodeCodex

Compress images, long prompts, conversations, and RAG documents locally via Ollama before consuming LLM tokens. Reduces token usage by 80% on images and 85% on text. Use proactively when: the user uploads images, processes long context, works with multiple RAG sources, or when approaching token limits.

24 1mo ago A 73 tokens

KathanModh259/latent-gate

Skill Claude CodeCodex

Compress images, long prompts, conversations, and RAG documents locally via Ollama before consuming LLM tokens. Reduces token usage by 80% on images and 85% on text. Use proactively when: the user uploads images, processes long context, works with multiple RAG sources, or when approaching token limits.

24 1mo ago A 73 tokens

cursorrules

06

KathanModh259/latent-gate

Cursor rule Cursor

Add this file to your project root as .cursorrules to make Cursor automatically use LatentGate for token compression.

24 1mo ago A 132 tokens

tokenfold AGENTS.md

07

snchimata/tokenfold

Instructions file CodexOpenCode

Instructions for snchimata/tokenfold, covering repository guidelines, working principles, repository map, implementation conventions and validation.

22 10d ago A 894 tokens original Apache-2.0

tokenwise

08

CodeShuX/tokenwise

Plugin Claude Code

Measurement-driven model router for Claude Code. Classifies every task by type — mechanical, execution, review, planning — and routes it automatically to Haiku, Sonnet, Opus, or Fable. Measures real dollars saved per session.

3 7d ago A tokens not measured original MIT

ab

09

CodeShuX/tokenwise

Skill Claude CodeCodex

Run an A/B test of the same task at multiple model tiers (Haiku, Sonnet, optionally Opus and Fable). Captures outputs, computes structural and semantic diffs, scores quality, writes a markdown comparison report. Use when the user wants to validate "is Haiku good enough for this task class?" or runs /tokenwise:ab " ".

3 7d ago A 76 tokens original MIT

install

10

CodeShuX/tokenwise

Skill Claude CodeCodex

Install TokenWise routing rules into the user's Claude Code config (CLAUDE.md and settings.json). Probes the user's Claude Code build for known routing bugs first, shows diffs, backs up originals, asks for per-file confirmation. Supports --guided (default), --manual (print copy-paste only), and --dry-run (preview…

3 7d ago B 95 tokens original MIT

report

11

CodeShuX/tokenwise

Skill Claude CodeCodex

Print a TokenWise session report — tokens per model, $ saved vs all-Opus baseline, reclassifications, quality flags. Reads from .tokenwise/log.ndjson filtered to the current session. Use when the user asks "how much did I save", "show me tokenwise stats", or runs /tokenwise:report.

3 7d ago A 68 tokens original MIT

cctx-optimizer

12

anil7948/cctx

MCP server Claude CodeCodexCursor +2

Reduce Claude Code token usage by 70-90% with cross-session memory and four-layer optimization: semantic codebase indexing, PostToolUse tool compression, turn summarization, and session consolidation. MCP server powered by Ollama — zero API calls, runs en. Runs locally from the cctx-optimizer npm package.

3 2mo ago A tokens not measured original MIT

Manas470/tokenoptim

Instructions file

Instructions for Manas470/tokenoptim, covering llm-tokenoptim — token optimization skill, output rules, auto-clarity rule and token budget awareness.

2 3mo ago A 256 tokens original MIT

ancient

14

Manas470/tokenoptim

Skill Claude CodeCodex

Stone tablet. Space precious. Every symbol costs.

2 3mo ago A 0 tokens original MIT

full

15

Manas470/tokenoptim

Skill Claude CodeCodex

Talk like caveman. Smart caveman — knows code, knows systems, knows math.

2 3mo ago A 0 tokens original MIT

standard

16

Manas470/tokenoptim

Skill Claude CodeCodex

Respond like a terse, senior engineer. Rules.

2 3mo ago A 0 tokens original MIT

iceage

17

pomagrenate/ice_age

Cursor rule Cursor

Iceage mode — terse communication, 75% fewer tokens, full technical accuracy.

2 13d ago A 190 tokens original Apache-2.0

pomagrenate/ice_age

Instructions file GitHub Copilot

Instructions for pomagrenate/ice_age: Respond terse like smart iceage. All technical substance stay. Only fluff die.

2 13d ago A 162 tokens original Apache-2.0

ice_age AGENTS.md

19

pomagrenate/ice_age

Instructions file CodexOpenCode

Instructions for pomagrenate/ice_age: @./skills/iceage/SKILL.md @./skills/iceage-commit/SKILL.md @./skills/iceage-review/SKILL.md @./skills/compress/SKILL.md.

2 13d ago A 46 tokens original Apache-2.0

ice_age CLAUDE.md

20

pomagrenate/ice_age

Instructions file

Instructions for pomagrenate/ice_age, covering claude.md — iceage, readme is a product artifact, project overview, file structure and what owns what and single source of truth files — edit only these.

2 13d ago B 3,141 tokens original Apache-2.0

compress

21

pomagrenate/ice_age

Skill Claude CodeCodex

Compress natural language memory files (CLAUDE.md, todos, preferences) into iceage format to save input tokens. Preserves all technical substance, code, URLs, and structure. Compressed version overwrites the original file. Human-readable backup saved as FILE.original.md. Trigger: /iceage:compress or "compress memory…

2 13d ago A 73 tokens copy · 84% Apache-2.0

iceage-batch

22

pomagrenate/ice_age

Skill Claude CodeCodex

Batch compress all natural language files in a directory using iceage compression. Parallel workers, progress tracking, dry-run mode. Use when user says "compress all files", "batch compress", "compress this directory", or invokes /iceage-batch.

2 13d ago A 55 tokens original Apache-2.0

iceage-export

23

pomagrenate/ice_age

Skill Claude CodeCodex

Exports the current chat session to a clean markdown file. Reconstructs the visible conversation from context — user messages, assistant responses, code blocks, tool output — formatted as readable .md. Use when user says "export chat", "save this conversation", "export to markdown", or invokes /iceage-export. Accepts…

2 13d ago A 73 tokens original Apache-2.0

tokensaver-mcp

24

pozii/tokensaver

MCP server Claude CodeCodexCursor

MCP server "tokensaver-mcp" as configured in pozii/tokensaver. Runs locally from the tokensaver-mcp Python package.

1 2mo ago A tokens not measured original Apache-2.0

At most 3 mods per repository are shown here — the rest are on their repository pages: