---
title: "Discussion on \"Clogmeter is live on Product Hunt: see what your agent instruction files cost\" | Hashnode"
description: "Discussion on \"Clogmeter is live on Product Hunt: see what your agent instruction files cost\". If you use Claude Code, Cursor or Codex, your agent re-reads the same instruction files on every turn: AGENTS.md, CLAUDE.md, Cursor rules, skills. They grow quietly. Rules get copied between files, an"
url: "https://hashnode.com/posts/clogmeter-is-live-on-product-hunt-see-what-your-agent-instruction-files-cost/6af2d0a15d456961fae8fe32"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Clogmeter is live on Product Hunt: see what your agent instruction files cost](https://clogmeter.hashnode.dev/clogmeter-is-live-on-product-hunt-see-what-your-agent-instruction-files-cost?utm_source=hashnode&utm_medium=feed)

If you use Claude Code, Cursor or Codex, your agent re-reads the same instruction files on every turn: AGENTS.md, CLAUDE.md, Cursor rules, skills. They grow quietly. Rules get copied between files, an

Nov 10 · 1 min read · clogmeter.hashnode.dev  
Read the article

-   [#ai](https://hashnode.com/tag/ai)
-   [#productivity](https://hashnode.com/tag/productivity)
-   [#claude](https://hashnode.com/tag/claude)
-   [#cursor](https://hashnode.com/tag/cursor)

## Written by

[Clogmeter](https://hashnode.com/@clogmeter)

@clogmeter

  
[View profile](https://hashnode.com/@clogmeter)

Free browser-only checker for token weight, duplicates and conflicts in AGENTS.md, CLAUDE.md and Cursor rules. https://clogmeter.vercel.app

Free browser-only checker for token weight, duplicates and conflicts in AGENTS.md, CLAUDE.md and Cursor rules. https://clogmeter.vercel.app

## Responses 2

Markdown supported · ⌘↵ to post

[Apps by Mo](https://hashnode.com/@appsbymo)

[Nov 10](https://hashnode.com/posts/clogmeter-is-live-on-product-hunt-see-what-your-agent-instruction-files-cost/6af2d0a15d456961fae8fe32/comment/6af318ab8d0998379daa40ba)

Good problem to measure. What helped me most with Claude Code was splitting the file that loads every session from everything else: my memory index is one line per note, and a note is only read when it is relevant. The index stays small while the notes grow. Does Clogmeter tell apart files read on every turn from ones loaded on demand?

[Clogmeter](https://hashnode.com/@clogmeter)

[Nov 10](https://hashnode.com/posts/clogmeter-is-live-on-product-hunt-see-what-your-agent-instruction-files-cost/6af2d0a15d456961fae8fe32/comment/6af318ab8d0998379daa40ba#reply-6af379592bc9340ddaa892a9)

Thanks Mo! Keeping the always-loaded file to a short index and reading notes on demand is exactly the kind of split Clogmeter helps you spot. Glad it's a problem worth measuring.
