What is AgentVibes CLAUDE.md?

Development instructions for AgentVibes, a system that gives coding agents a spoken voice through hooks, an MCP server, and audio providers.

Use them when developing or maintaining AgentVibes features such as speech playback, remote receivers, configuration, and tests.

They document the project structure and rules that developers need to follow when changing the system.

AgentVibes CLAUDE.md — Instructions file for coding agents by paulpreibisch: Development instructions for AgentVibes, a system that gives co

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/paulpreibisch/agentvibes/claude-md

Clone the repo

git clone --depth 1 https://github.com/paulpreibisch/AgentVibes

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.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/paulpreibisch/agentvibes/claude-md.svg)](https://agentmods.dev/instructions/paulpreibisch/agentvibes/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/paulpreibisch/agentvibes/claude-md"><img src="https://agentmods.dev/badge/instructions/paulpreibisch/agentvibes/claude-md.svg" alt="Measured on agentmods" height="20"></a>

Per session 1,947 This file is loaded in full into every session.

When invoked 1,947 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.

Model	Per session	Once invoked
Fable 5	$0.01947	$0.01947
Opus 5	$0.00974	$0.00974
Sonnet 5	$0.00389	$0.00389
Haiku 4.5	$0.00195	$0.00195

Measured yesterday against content hash 1cc7b444b7b6, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

AgentVibes 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 yesterday.

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 · 88 lines

How it starts

The opening of the file, as written. The whole thing — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AgentVibes — Development Guidelines

Updated: 2026-07-06 (v5.12.x era)

What this is

AgentVibes (agentvibes on npm) gives AI coding agents a spoken voice: Claude Code hooks + MCP server + interactive TUI, with Piper/Kokoro/ElevenLabs providers, background music, per-agent BMAD voices, a remote SSH receiver, and avatar rendering. Node ESM core with parallel bash and PowerShell runtimes and a Python MCP server.

Repo map

Path	What it is
`src/installer.js` + `src/installer/`	npx installer (manifest-based, non-destructive)
`src/console/`	blessed TUI (tabs, modals, widgets)
`src/services/`	config-service, provider services, utterance-resolver.js (single source of truth for playback plans)
`.claude/hooks/` + `.claude/hooks-windows/`	Shipped product code — bash/PS1 runtime (`play-tts.sh`, `audio-processor.sh`, `bmad-speak.sh`, …) installed into user projects
`mcp-server/`	Python MCP server (`server.py`, pytest suite)
`templates/`	Remote receiver scripts (`agentvibes-receiver.sh`/`.ps1`)
`bin/`	CLI entry points (`agentvibes.js`, `resolve-utterance.js`, `mcp-server.js`)
`test/unit/`	node:test `.test.js` + bats `.bats` suites
`docs/implementation-artifacts/`	BMAD story files + `sprint-status.yaml`
`docs/architecture/`	Provider/system architecture docs

Everything under the repo's .claude/ (hooks, personalities, commands, config) is the npm package payload. Editing a hook here is a product change: it must pass tests, keep bash/PowerShell parity, and reach real installs via the installer (manifestSafeCopy) — not just the dev checkout. The remote receiver additionally needs an in-place redeploy to ~/.agentvibes/ + watcher restart.

Architecture invariants (hard-won — do not regress)

The utterance resolver is the single source of truth. src/services/utterance-resolver.js reads all config once and emits a flat plan (voice, engine, transport, effects, music, speed, volume). Players consume the plan; they must NOT re-read config files directly. New playback behavior goes in the resolver + precedence table, not in a player script.
Volume defaults are 20% (0.20), never 70%. Regression-tested; covered in bash, PS, and JS.
Never select TTS output by "most recent file." Providers emit an AV_OUTPUT: sentinel on stdout; capture the exact filename and fail loudly on synthesis failure. The ls -t heuristic is retired.
Kokoro voices force the kokoro engine (an af_*/am_* voice with a piper ENGINE column = silence). The resolver encodes this; keep it.
Bash/PowerShell parity. Any change to play-tts.sh needs the mirror change in play-tts.ps1 (and vice versa), plus the contract-test matrix (AVI-S8.7) kept green. Note the historical env-flag split: bash reads AGENTVIBES_NO_PLAYBACK, PS reads AGENTVIBES_NO_PLAY — set both.
Receiver settings are keyed by project (TTS --project-dir basename namespaces scene/avatar/voice). Don't wipe glb-scenes.json — it holds calibration.
SSH senders must never force -p 22 — it overrides ~/.ssh/config alias ports and silently drops TTS.

Read the full file on GitHub · 88 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.

yesterday First seen · 88 lines · 1,947 tokens per session scan A 1cc7b444b7b6

Subscribe to this mod's changes

AgentVibes CLAUDE.md is an instructions file published in the GitHub repository paulpreibisch/AgentVibes (153 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 1,947 tokens to every session, about $0.0097 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 2026-09-03.

Creative and media ai-agents collaborative-coding docker github-projects team-generator themed-agents

AgentVibes CLAUDE.md

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

AgentVibes — Development Guidelines

What this is

Repo map

Architecture invariants (hard-won — do not regress)

What this file has done since we first saw it

Other instructions, from other repositories

appwrite AGENTS.md

SForum AGENTS.md

selfhost-ai CLAUDE.md

redd-archiver CLAUDE.md

homelab-monitor CLAUDE.md

yoloai CLAUDE.md