sflow CLAUDE.md

sflow CLAUDE.md is an instructions file for coding agents from daniel-carreon/sflow. It costs 4,518 tokens per session, scanned A, original, MIT.

A historical development guide for SFlow v1/v2, a macOS voice-to-text app that records speech through keyboard shortcuts, sends it to the Groq Whisper transcription service, and pastes the text where the cursor is.

In plain words
What is it for?
Use it only as a reference for the older Python app's setup and behavior; new SFlow work should use the sflow-next project described in the guide.
Why use it?
It prevents developers from working on an old Python version when current SFlow development belongs in the separate Swift-based v3 project.

Instructions file

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/daniel-carreon/sflow/claude-md
Clone the repo
git clone --depth 1 https://github.com/daniel-carreon/sflow

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for sflow CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/daniel-carreon/sflow/claude-md.svg)](https://agentmods.dev/instructions/daniel-carreon/sflow/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/daniel-carreon/sflow/claude-md"><img src="https://agentmods.dev/badge/instructions/daniel-carreon/sflow/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 4,518 This file is loaded in full into every session.
When invoked 4,518 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.04518 $0.04518
Opus 5 $0.02259 $0.02259
Sonnet 5 $0.00904 $0.00904
Haiku 4.5 $0.00452 $0.00452

Measured 3d ago against content hash 34a999232355, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

sflow CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

CLAUDE.md · 330 lines

How it starts

The opening of the file, as written. The whole thing — 330 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md — SFlow Development Instructions

⚠️ ESTE ES SFlow v1/v2 (Python) — SUPERADO POR v3

La versión viva es SFlow v3, un rebuild nativo en Swift: repo daniel-carreon/sflow-next · local ~/Developer/software/sflow-next/. Es la que está instalada en /Applications/SFlow.app (aquí quedó SFlow-v1-backup.app).

Este repo se conserva SOLO como referencia histórica. No lo desarrolles. Si te pidieron trabajar en "SFlow", es casi seguro que quieren sflow-next. v3 comparte el mismo data dir (~/Library/Application Support/SFlow/) y los mismos IDs de modelo — la migración v1→v3 fue cero copia.

What is SFlow?

SFlow is a macOS voice-to-text desktop tool that replaces Wispr Flow ($15/month). It captures audio via global hotkeys, transcribes using Groq Whisper API (~$0.02/hour), and auto-pastes text wherever the cursor is. It includes a floating pill UI overlay, real-time audio visualization, SQLite history, and a web dashboard.

Quick Start (Dev Mode)

# 1. Install system dependency
brew install portaudio

# 2. Create virtual environment
python3 -m venv venv
source venv/bin/activate

# 3. Install Python dependencies
pip install -r requirements.txt

# 4. Set up environment
cp .env.example .env
# Edit .env and add your GROQ_API_KEY (get one at https://console.groq.com/keys)

# 5. Run
python3 main.py

Build Desktop App (.app bundle)

# Recommended: full install dance (build + ditto + kill old proc + relaunch + open Accessibility panel)
bash install.sh

# Or just build the bundle without installing
bash build.sh

install.sh exists because build.sh alone leaves you with a broken paste — see "Critical: ad-hoc rebuild → silent Accessibility revocation" below.

The .app bundle is self-contained (~478MB — INCLUYE el stack MLX: mlx-whisper + parakeet-mlx + librosa/numba/scipy + los Metal .metallib). No Python, no venv, no terminal. On first launch, if no API key exists in ~/Library/Application Support/SFlow/.env, a dialog asks for it. Menu-bar app.

Read the full file on GitHub · 330 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 330 lines · 4,518 tokens per session scan A 34a999232355

Subscribe to this mod's changes

sflow CLAUDE.md is an instructions file published in the GitHub repository daniel-carreon/sflow (22 stars, last pushed 17d ago), licensed MIT. It adds 4,518 tokens to every session, about $0.0226 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.