Borrowing it
Nothing to install: this file belongs to T0mSIlver/localvoxtral. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/T0mSIlver/localvoxtral/main/AGENTS.mdgit clone --depth 1 https://github.com/T0mSIlver/localvoxtralWrote 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.
[](https://agentmods.dev/instructions/t0msilver/localvoxtral/agents-md)<a href="https://agentmods.dev/instructions/t0msilver/localvoxtral/agents-md"><img src="https://agentmods.dev/badge/instructions/t0msilver/localvoxtral/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.
<a href="https://agentmods.dev/instructions/t0msilver/localvoxtral/agents-md"><img src="https://agentmods.dev/badge/instructions/t0msilver/localvoxtral/agents-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.04110 | $0.04110 |
| Opus 5.5 | $0.01644 | $0.01644 |
| Sonnet 5.5 | $0.00822 | $0.00822 |
| Haiku 4.5 | $0.00411 | $0.00411 |
Grade A, and why
localvoxtral AGENTS.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.
How it starts
The opening of the file, as written. The whole thing — 267 lines — stays where its author put it; the contents beside it link to each section on GitHub.
localvoxtral agent guide
Native macOS menu bar app for realtime dictation: Swift 6.2 strict
concurrency, SwiftPM, macOS 15+. It streams mic audio to an OpenAI
Realtime-compatible backend (the bundled localvoxtral-speechd helper, or any
compatible server in External URL mode) and inserts text into the focused app,
either live (Live Auto-Paste) or through an overlay committed on stop (Overlay
Buffer, with LLM polishing). Subsystem map: docs/architecture.md. It has
daily users, so nothing ships on "it compiles".
Build and test
The app only compiles on macOS. From Linux, ./scripts/remote-build.sh rsyncs
the working tree to the Mac build host and runs the toolchain there; no commit
needed. Set the host once per clone: git config localvoxtral.buildhost <ssh-destination>. The script's header lists every verb.
Two self-hosted Macs: the Mac Mini (8 GB, label mini) is the build host and
runs PR CI, UI Smoke, the screenshot lanes, record-demo and the release; the
owner's MacBook (labels self-hosted, macOS, ARM64) keeps the evals and crash
logs. Spend them only where nothing else can do the job:
- Linux first.
localvoxtralCore(the Foundation-only pieces) builds and tests here:./scripts/core-tests-linux.sh(Swift 6.2;SWIFT=names the toolchain; a cloud session gets it fromscripts/linux/cloud-setup.sh, the environment's setup script), and so doscripts/ci/test-*.sh. The app re-exports core, so a core declaration the app uses needspackageaccess. On Linux a@MainActortest class with noasynctest aborts the whole test process at discovery; make oneasync. - On the Mac, run only the suites your change touches:
remote-build.sh test --filter <Suite>, the flag repeated per suite (the host's SSH gate refuses|). Never run the full suite there: the PR's hostedbuild-testruns it on every push, drafts included. The script refuses atestwith no--filterand the live verbs below unlessLV_ALLOW_HEAVY_MAC_RUN=1is set; set it only when a rule here requires. - Live lanes and evals (
integration*,eval-e2e,eval-term-recall) run minutes to half an hour of inference. Run one only when a rule below requires it, once, on the final diff. When the PR's CI lane runs it, don't run it by hand too. Iterate on polish against a frozen eval log withscripts/ablate-agent-eval.py. - Run
./scripts/mac-health.shbefore remote work. A sleeping Mac makes rsync hang instead of fail. - Never pipe
remote-build.shthrough grep: a crash eats the failing test's name. The full output is in.build/last-remote.log, and the local Linux part's in.build/last-linux.log. - An interrupted run can leave a stale SwiftPM lock. Switch to a fresh
LV_BUILD_DIRinstead of debugging it. Never hand-clean~/workon the Mac; abandoned build dirs are garbage-collected. - Only
package_app.shproduces working Metal kernels for the two MLX helpers. ReadPolishHelper/AGENTS.mdorSpeechHelper/AGENTS.mdbefore touching either.
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.
- 3d ago Changed · +2 tokens per session d31902e99989
- 13d ago Changed · +2 lines · +51 tokens per session 2c6d9e7a0620
- 16d ago Changed · +1 lines · +31 tokens per session 99b55b5e8489
- 17d ago Changed · +4 lines · +65 tokens per session 3c185e1acc6c
- 18d ago Changed b20cdeea9fd5
- 19d ago Changed · -2 lines · -37 tokens per session 4e6f21ee651a
- 20d ago Changed · +9 lines · +141 tokens per session f8b06a72a66e
- 21d ago Changed · +1 lines · +12 tokens per session ed94ed557952
- 22d ago Changed · -4 lines · -104 tokens per session 37beeb8181e4
- 23d ago First seen · 256 lines · 3,949 tokens per session scan A 832e698f6f92
localvoxtral AGENTS.md is an instructions file published in the GitHub repository T0mSIlver/localvoxtral (59 stars, last pushed yesterday), licensed MIT. It adds 4,110 tokens to every session, about $0.0164 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.
Other instructions, from other repositories
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).
deno copilot-instructions.md
Copilot instructions for denoland/deno, covering deno development guide for github copilot, network access, high level overview, key directories and key files to understand first.