agentmods.dev

FluidAudio: Instructions file for Claude Code

CLAUDE.md

FluidAudio CLAUDE.md is an instructions file for Claude Code from FluidInference/FluidAudio. It costs 2,742 tokens per session, scanned A, original, Apache-2.0.

Development instructions for FluidAudio, a Swift framework for local audio processing on Apple devices. It covers speaker identification, speech-to-text, and detecting when speech is present, using Core ML models.

In plain words
What is it for?
Developing or testing FluidAudio's audio features, handling thread safety with Swift's proper synchronization tools, and working with real Core ML models and model uploads.
Why use it?
It sets safety and correctness rules, including avoiding unsafe concurrency shortcuts and fake models or test audio.

Instructions file for Claude Code

Written for Claude Code: the file is CLAUDE.md. Also seen: mentions CLAUDE.md; mentions Claude Code.

This is FluidInference/FluidAudio's own configuration. It tells Claude Code how to work on FluidAudio itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything FluidAudio configures →

Needs its repository: it reads a path above its own folder, which exists only inside the repository. The line is git worktree add ../FluidAudio-<task> -b <branch> origin/main.

not rated 3.3krepo +29 changed 23d ago A scan 2,742 tokens original Apache-2.0
View source ↗ FluidInference/FluidAudio docs.fluidinference.com
About the project

FluidAudio is a Swift SDK that runs local audio AI models on Apple devices for speech recognition, text-to-speech, voice activity detection, and identifying who is speaking. It is intended for developers building iOS and macOS apps with low-latency audio processing. The catalogue add-ons provide workflows for using the SDK.

FluidInference/FluidAudio · 3,265 stars · on GitHub · docs.fluidinference.com

Reuse

Borrowing it

Nothing to install: this file belongs to FluidInference/FluidAudio. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/FluidInference/FluidAudio/main/CLAUDE.md
Clone the repo
git clone --depth 1 https://github.com/FluidInference/FluidAudio

Made for: Claude Code.

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 FluidAudio CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/fluidinference/fluidaudio/claude-md/github.svg)](https://agentmods.dev/instructions/fluidinference/fluidaudio/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/fluidinference/fluidaudio/claude-md"><img src="https://agentmods.dev/badge/instructions/fluidinference/fluidaudio/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for FluidAudio CLAUDE.md

Your own site · 80×15
<a href="https://agentmods.dev/instructions/fluidinference/fluidaudio/claude-md"><img src="https://agentmods.dev/badge/instructions/fluidinference/fluidaudio/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 2,742 This file is loaded in full into every session.
When invoked 2,742 The same file — it is already loaded in full.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.02742 $0.02742
Opus 5.5 $0.01097 $0.01097
Sonnet 5.5 $0.00548 $0.00548
Haiku 4.5 $0.00274 $0.00274

Measured 23d ago against content hash f43efa929438, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

FluidAudio 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 23d 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 · 207 lines

How it starts

The opening of the file, as written. The whole thing — 207 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

FluidAudio is a Swift framework for local, low-latency audio processing on Apple platforms. It provides speaker diarization, automatic speech recognition (ASR), and voice activity detection (VAD) through open-source models converted to Core ML.

Critical Development Rules

NEVER USE @unchecked Sendable

  • Always implement thread-safe code with proper synchronization
  • Use actors, @MainActor, or proper locking mechanisms instead
  • If you encounter Sendable conformance issues, fix them properly

NEVER CREATE DUMMY MODELS OR SYNTHETIC DATA

  • Do not create dummy, mock, or fake models for testing or development
  • Do not generate synthetic audio data for testing
  • Always use the actual models required by the code
  • If model authentication is required, inform the user rather than creating dummy versions

HUGGINGFACE UPLOADS

  • Uploading models, datasets, and files to HuggingFace is allowed
  • Creating HuggingFace repos is allowed
  • Confirm the target repo with the user before uploading
  • Traceability (issue ↔ code ↔ HF): every model fix/upload must be linkable both ways
    • Ship fixed models under a new name (Foo_v2.mlmodelc) next to the original; never overwrite in place
    • HF commit message names the issue: ... (FluidAudio #NNN); pass --commit-description with the mobius/FluidAudio PR links
    • mobius and FluidAudio PR bodies link the issue, each other, and the full HF commit URL
    • Add a row to the HF model card's ## Changelog table (date, files, change, issue, mobius PR, FluidAudio PR, HF commit); create the section if missing
    • The Swift ModelNames comment for a renamed file cites the issue number

MODEL OPERATIONS - CONSULT BEFORE IMPLEMENTING

  • When asked to merge, convert, or modify models:
    • If it seems impossible or there are significant objections, consult the user first
    • If they say proceed, do it immediately without further objections
  • Do not create placeholder models or implement alternatives without asking

Read the full file on GitHub · 207 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. 23d ago Changed · +28 tokens per session f43efa929438
  2. 1mo ago Changed · +6 lines · +171 tokens per session 255a9f6f7d7e
  3. 2mo ago First seen · 201 lines · 2,543 tokens per session scan A 029c5322c0b9

Subscribe to this mod's changes

FluidAudio CLAUDE.md is an instructions file published in the GitHub repository FluidInference/FluidAudio (3,265 stars, last pushed 4d ago), licensed Apache-2.0. It adds 2,742 tokens to every session, about $0.0110 per session on Opus 5.5. 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.

Kotlin and SwiftLanguagesMobilecoremliosmacosspeaker-diarizationspeaker-embeddingspeaker-identification
Related

FluidAudio CLAUDE.md compared with instructions from other repositories

FluidAudio CLAUDE.md next to 6 instructions from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
FluidAudio CLAUDE.mdFluidInference/FluidAudioInstructions file3.3k4d agoApache-2.0A2,742Claude Code
vscode buildNext.instructions.mdmicrosoft/vscodeInstructions file195ktodayMITA9,719GitHub Copilot
next.js AGENTS.mdvercel/next.jsInstructions file143ktodayMITA7,461Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchainInstructions file149ktodayMITA4,680Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscodeInstructions file195ktodayMITA5,516GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscodeInstructions file195ktodayMITA3,093GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harnessInstructions file274k2d agoMITA4,313Codex, OpenCode