rtk-ai/rtk

cluade.md bloat

Open

#598 aberto em 15 de mar. de 2026

Ver no GitHub
 (2 comments) (3 reactions) (0 assignees)Rust (2.914 forks)batch import
area:docsdocumentationeffort-smallenhancementgood first issuehelp wantedpriority:low

Métricas do repositório

Stars
 (48.085 stars)
Métricas de merge de PR
 (Mesclagem média 11d 1h) (45 fundiu PRs em 30d)

Description

First of all, well done and thanks! Second, for a project aiming to improve token spend - you bloat the claude.md in a terrible way without telling us. Why do we need there info on token consumption? Why the word Cargo needs to be added to my context for a js project? etc. It's like an hidden ad, not cool.

This is my summarized lines. I suggest you find the best 3-liner instead of the current bloat.

<!-- rtk-instructions -->
# RTK - Token-Optimized CLI Proxy (60-90% savings)
Always prefix ALL commands with `rtk` (e.g., `rtk git status`, `rtk pnpm install`, `rtk docker ps`). Works with git, gh, pnpm/npm/npx, docker, kubectl, curl, cargo, tsc, vitest, and more.
If RTK has a filter it compresses output; if not it passes through unchanged — so it's always safe to use. In command chains: `rtk git add . && rtk git commit -m "msg"`.
Meta commands: `rtk gain` (savings stats), `rtk discover` (find missed usage), `rtk proxy <cmd>` (bypass filtering for debug).
<!-- /rtk-instructions -->

And thanks again

Guia do colaborador