What is voice_clone_lab AGENTS.md?

A repository instruction file for Voice Clone Lab, a tool that fine-tunes a speech model on a person’s voice and generates audio from it.

Use it when changing the project so its configuration, vendored dependency, runtime data, command-line tool, user interface, and CPU-only tests are handled consistently.

It gives coding agents the project’s layout, configuration rules, testing practices, and safety constraints for handling voice data.

How tetsuo-ai/voice_clone_lab configures Codex and OpenCode

Reuse

Borrowing it

Nothing to install: this file belongs to tetsuo-ai/voice_clone_lab. 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/tetsuo-ai/voice_clone_lab/main/AGENTS.md

Clone the repo

git clone --depth 1 https://github.com/tetsuo-ai/voice_clone_lab

Made for: Codex, OpenCode.

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/tetsuo-ai/voice_clone_lab/agents-md/github.svg)](https://agentmods.dev/instructions/tetsuo-ai/voice_clone_lab/agents-md)

Your own site

<a href="https://agentmods.dev/instructions/tetsuo-ai/voice_clone_lab/agents-md"><img src="https://agentmods.dev/badge/instructions/tetsuo-ai/voice_clone_lab/agents-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.

Your own site · 80×15

<a href="https://agentmods.dev/instructions/tetsuo-ai/voice_clone_lab/agents-md"><img src="https://agentmods.dev/badge/instructions/tetsuo-ai/voice_clone_lab/agents-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 895 This file is loaded in full into every session.

When invoked 895 The same file — it is already loaded in full.

Security scan C 1 finding. 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.

Model	Per session	Once invoked
Fable 5.1	$0.00895	$0.00895
Opus 5.5	$0.00358	$0.00358
Sonnet 5.5	$0.00179	$0.00179
Haiku 4.5	$0.00089	$0.00089

Measured 2mo ago against content hash abc71b57da61, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade C, and why

voice_clone_lab AGENTS.md scanned grade C with 1 finding 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 2mo 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.

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

No unconditional `ffmpeg -y`, no unguarded `rm -rf`/`shutil.rmtree`. The one

AGENTS.md · 69 lines

How it starts

The opening of the file, as written. The whole thing — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

Guidance for agents working in this repo.

What this is

Voice Clone Lab: an installable pipeline (src/voice_clone_lab/, console script vcl) for fine-tuning Qwen3-TTS-12Hz-1.7B-Base on a user's own voice and generating speech. CLI (cli.py) + Gradio UI (ui.py) over shared modules.

Layout

src/voice_clone_lab/ — the package. config.py (dataclass config + per-voice path derivation), utils.py (IO + overwrite safety), audio.py, transcribe.py, dataset.py, train.py, generate.py, sources.py (yt-dlp URL downloads), system.py, cli.py, ui.py.
config/default.yaml — the ONLY config source. Every key is used; if you add a setting, add it here AND to the dataclasses in config.py.
patches/sft_12hz.patch — local patch to the vendored training script. See "Vendor rule" below.
third_party/Qwen3-TTS/ — vendored upstream clone (gitignored, recreated by vcl setup).
data/, outputs/, models/, attic/ — gitignored runtime dirs. data/ and outputs/ contain biometric voice data: never commit, never upload.
tests/ — pytest, CPU-only (synthetic audio, tmp_path, monkeypatching).

Conventions (keep them when editing)

Overwrite safety: every step that writes files must refuse to overwrite unless force=True, via utils.refuse_overwrite / refuse_any_existing. No unconditional ffmpeg -y, no unguarded rm -rf/shutil.rmtree. The one destructive path (train --force) must stay guarded to outputs/checkpoints/ only.
Relative paths in artifacts: files under data/ that reference audio (chunk metadata, transcripts, dataset rows) store project-relative paths via config.rel_to_root; readers resolve with utils.resolve_artifact_path (accepts legacy absolute paths too).
Config idiom: values come from the Config dataclasses; function parameters override (None = use config). No config.get(...).get(...) chains, no second YAML reader.
Speaker-derived paths: never hardcode a voice name or a data path; use cfg.paths_for(speaker).
Lazy heavy imports: torch, qwen_tts, soundfile, ASR backends, webrtcvad, noisereduce import inside functions so vcl stays fast and CPU-only tests pass.
User-facing errors: raise SystemExit(clear message), not tracebacks.
Ported behavior is faithful to the original scripts — don't change filter chains, thresholds, or scoring formulas without a reason and a README note.

Read the full file on GitHub · 69 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.

2mo ago First seen · 69 lines · 895 tokens per session scan C abc71b57da61

Subscribe to this mod's changes

voice_clone_lab AGENTS.md is an instructions file published in the GitHub repository tetsuo-ai/voice_clone_lab (155 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 895 tokens to every session, about $0.0036 per session on Opus 5.5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Creative and media gradio qwen text-to-speech tts voice-cloning

voice_clone_lab AGENTS.md next to 6 instructions from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
voice_clone_lab AGENTS.mdtetsuo-ai/voice_clone_lab	Instructions file	155	3mo ago	Apache-2.0	C	895	Codex, OpenCode
vscode buildNext.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	9,719	GitHub Copilot
vscode oss-third-party-notices.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	5,516	GitHub Copilot
next.js AGENTS.mdvercel/next.js	Instructions file	143k	today	MIT	A	7,461	Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchain	Instructions file	149k	today	MIT	A	4,680	Codex, OpenCode
vscode copilot-instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	3,093	GitHub Copilot
vscode model-prompts.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	2,022	GitHub Copilot

voice_clone_lab: Instructions file for Codex