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 skills add sonichi/sutando --skill model-switchgit clone --depth 1 https://github.com/sonichi/sutandoWrote 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/skills/sonichi/sutando/model-switch)<a href="https://agentmods.dev/skills/sonichi/sutando/model-switch"><img src="https://agentmods.dev/badge/skills/sonichi/sutando/model-switch/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/skills/sonichi/sutando/model-switch"><img src="https://agentmods.dev/badge/skills/sonichi/sutando/model-switch.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00056 | $0.00717 |
| Opus 5 | $0.00028 | $0.00358 |
| Sonnet 5 | $0.00011 | $0.00143 |
| Haiku 4.5 | $0.00006 | $0.00072 |
Grade A, and why
model-switch 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 4d 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
model-switch
scripts/switch-model.sh <model> [--dry-run] [--confirm] [--accept-timeout S] — <model> is one of the
CLI's aliases (default|opus|sonnet|haiku|fable) or a claude-* id with an optional [1m] tag;
anything else is refused before any write (exit 2).
Order, under one lock per brain (<workspace>/state/.model-switch.lock, held across preflight, send,
observation and record so overlapping switches linearize): (1) preflight through the shared sender
(scripts/tmux-send-line.sh --dry-run --refuse-if-pending) — refuse on a Codex runtime (exit 4) or
when the core's input box carries text (exit 5), nothing written; exit 3 when no live pane exists;
(2) count the Set model to lines already on screen (scripts/pane-observe.sh --count), so a stale
acceptance from an earlier switch cannot pass as this one; (3) /model <model> + Enter through the
shared sender; (4) wait for the CLI to accept: a NEW Set model to line is the switch. A warm
core (cached conversation) first shows "Yes, switch / No, go back"; with --confirm — pass it on
an owner instruction, the flag is the authorization — the helper answers Enter once and waits again;
without it the dialog is cancelled (Escape) and the script exits 6 with nothing recorded. No
acceptance within --accept-timeout (default 20 s) is exit 8, nothing recorded. (5) Only then record
<workspace>/state/model-switch.json {model, previous, previous_source, accepted, confirmed, ts, by, settings_read} — previous from the runtime's settings.json, else the last record (the CLI does not
persist every /model pick, so the file can lag the live session), else null.
The script never writes settings.json: Claude Code's own /model persists the choice there
(measured 2026-09-04: the key changed hours into a session with no repo writer). Measured the same
day on a live core: the send alone exited 0 while the CLI sat at the confirm dialog and the model had
not changed — which is why the send is treated as initiation and the record follows acceptance.
Defaults come from the runtime descriptor (sutando-config.sh runtime: brain, socket,
session); --brain/--socket/--session/--descriptor-file and SUTANDO_TMUX_* are explicit
overrides.
Python comes from sutando-config.sh python-bin, never bare python3. The core-side entry
scripts/switch-model.sh only execs this script. An owner message like "switch model to opus" is
handled by running it; the dashboard's Quota tile shows the model the proxy then sees.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 4d ago First seen · 39 lines · 56 tokens per session scan A 9559005f70c6
model-switch is a skill published in the GitHub repository sonichi/sutando (394 stars, last pushed today), licensed MIT. It adds 56 tokens to every session and 717 once invoked, about $0.0003 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 2026-09-05.
Other skills, from other repositories
contacts
Manage contacts, communication channels, access control, and invite links.
gmail
Manage Gmail email — drafting, sending, organizing, filters, vacation replies, and inbox analysis.
app-builder
Build and edit small, personal visual tools and artifacts — dashboards, trackers, calculators, data visualizations, charts, simple landing pages, and slide decks the user wants for THEMSELVES. This is the right skill whenever the user asks to "visualize this," "make a chart," or "build an artifact" for their own use…
assistant-migration
Migrate from ChatGPT, Claude, OpenClaw, Hermes, Manus, and other AI assistants into Vellum by inspecting their data exports, conversation archives, files, prompts, custom instructions, memory, saved memories, tools, GPTs, workflows, integrations, and relationships, then mapping as much as safely possible into Vellum…
vellum-memory-v3-migration
One-time migration of an existing memory-v2 concept corpus into the memory-v3 section-grain "wiki" — topical articles with a stand-alone lead and queryable sections — with loss-proof staging, assistant-reviewed authoring, and a retrieval-eval gate before cutover.
media-processing
Ingest and process media files (video, audio, image).