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.
npx agentmods add commands/punt-labs/vox/modelgit clone --depth 1 https://github.com/punt-labs/voxWhat 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 | $0.00011 | $0.00401 |
| Opus 5 | $0.00005 | $0.00200 |
| Sonnet 5 | $0.00002 | $0.00080 |
| Haiku 4.5 | $0.00001 | $0.00040 |
Grade A, and why
model 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 2d 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.
What it actually says
/vox:model command
Switch the TTS model for the current provider mid-session. No argument opens
a picker; a name (full or shorthand) writes it to the session and to
.punt-labs/vox/vox.md.
Usage
/vox:model— pick from the models the current provider offers/vox:model v3— set model toeleven_v3(elevenlabs shorthand)/vox:model eleven_flash_v2_5— set by full name
Implementation
First normalize $ARGUMENTS: trim whitespace. An empty string counts as no
argument.
- (no argument):
- Call the
modelMCP tool with no arguments to get{"available": [...], "current": "..."}for the currently selected provider. - If
availableis empty, tell the user"No models to pick for this provider."and stop. - Otherwise call
AskUserQuestionwith one question:question:"Which model?"header:"Model"multiSelect: falseoptions: one per name inavailable; the current selection'sdescriptionis"(current)", others"".
- On the user's pick, call
modelagain withname=<pick>. No text output — the panel confirms.
- Call the
<name>: Call themodelMCP tool withname="<name>"directly. Elevenlabs shorthand (v3,flash,turbo,multilingual) resolves server-side. No text output.
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.
- 2d ago First seen · 40 lines · 11 tokens per session scan A 827427fd2a14
model is a command published in the GitHub repository punt-labs/vox (3 stars, last pushed 2d ago), licensed MIT. It adds 11 tokens to every session and 401 once invoked, about $0.0001 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 2026-08-31.
Other commands, from other repositories
code2model-dev
Create a Z specification for stateful entities in a system.
b-animate-dev
Animate and model-check a B machine with probcli.
speech-stop
Stop any audio Claude Speech is currently playing or has queued for this session.
check
Type-check a Z specification with fuzz.
code2model
Create a Z specification for stateful entities in a system.
b-animate
Animate and model-check a B machine with probcli.