gmickel

Flowmeter Public — Development skill for Claude Code

Development community

Local, privacy-first token, cost, cache, and trace analysis for Cursor, Claude Code, and Codex.

How to install Flowmeter Public

This entry records only its repository, not the path inside it, so there is no exact command to give. Open gmickel/flowmeter-public and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Flowmeter Public does

Local, privacy-first token, cost, cache, and trace analysis for Cursor, Claude Code, and Codex. Optional Flow-Next attribution.

Alternatives in Development

JWT Scan — JWT attack toolkit (offline) — alg:none forgery, RS256→HS256 algorithm confusion, weak-secret crack, static cl 4.5k ★
Benjamin Plus Skill — Benjamin-Plus: a measured token-efficiency skill for coding agents (−17.9% cost median, quality unchanged) 272 ★
Dexscreener Unofficial CLI + MCP + Skills — You are a token scanning specialist using the Dexscreener Unofficial CLI (not affiliated with or endorsed by D 213 ★

README

flowmeter

[![CI](https://github.com/gmickel/flowmeter-public/actions/workflows/ci.yml/badge.svg)](https://github.com/gmickel/flowmeter-public/actions/workflows/ci.yml) [![Python 3.11+](https://img.shields.io/badge/python-3.11%2B-3776AB.svg)](https://www.python.org/) [![MIT license](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)

Flowmeter reads local AI coding traces and turns them into token, cost, cache, workflow, and coverage reports. It supports Cursor, Claude Code, and Codex. Optional [Flow-Next](https://github.com/gmickel/flow-next) integration adds workflow attribution from tasks, skills, stages, and Git-bound receipts. Flowmeter's basic reports work without it.

Grok models used through Cursor appear under the Cursor host with their Grok model name. Native Grok Build session ingestion is planned but is not shipped in this release.

Everything runs on your machine. The default database is `~/.flowmeter/flowmeter.duckdb`.

Quick start

You need Python 3.11 or newer and [uv](https://docs.astral.sh/uv/).

git clone https://github.com/gmickel/flowmeter-public.git flowmeter
cd flowmeter
uv tool install .

flowmeter doctor
flowmeter ingest
flowmeter summary
flowmeter dashboard

Open [http://127.0.0.1:8787/](http://127.0.0.1:8787/). The first ingest reads supported history already on the machine. Flow-Next, hooks, a trace salt, and `watch` are optional.

Flowmeter is currently a `0.1.0` tester build. Expect rough edges. The evidence model is deliberately conservative: missing data stays missing, partial scans stay partial, and unknown model prices stay unpriced.

What is required

`flowmeter ingest` is the only step required to read the supported history already on your machine. Everything else improves coverage or presentation.