/ag2space-voice
s.space voice surface — binds the room the s.space client announces to the voice session's origin, so work delegated by voice in a room is answered in that room, and moves the client by voice (navigate_ui).
Claude Code
$ npx -y skills add sonichi/sutando --skill ag2space-voice --agent claude-code
Codex
$ npx skills add sonichi/sutando --skill ag2space-voice -a codex -y
Cursor
$ npx skills add sonichi/sutando --skill ag2space-voice -a cursor -y
Gemini CLI
$ npx skills add sonichi/sutando --skill ag2space-voice -a gemini-cli -y
GitHub Copilot
$ npx skills add sonichi/sutando --skill ag2space-voice -a github-copilot -y
OpenCode
$ npx skills add sonichi/sutando --skill ag2space-voice -a opencode -y
Windsurf
$ npx skills add sonichi/sutando --skill ag2space-voice -a windsurf -y
SKILL.md
ag2space-voice.SKILL.mdname: ag2space-voice
description: "s.space voice surface — binds the room the s.space client announces to the voice session's origin, so work delegated by voice in a room is answered in that room, and moves the client by voice (navigate_ui)."
when_to_use: "Loaded automatically at voice-agent startup as a manifest skill. Not slash-invoked. Only an s.space client sends the frames it handles; on any other install it stays idle."
s.space voice
Optional plugin for installs that talk to an s.space client (desktop app or webapp). The host engine carries a product-neutral *voice session origin* (`VoiceSessionOrigin`, `src/task-bridge.ts`) and offers client frames to skills (`setup(ctx)`, see [`skills/MANIFEST.md`](../MANIFEST.md)). This skill owns everything s.space specific on top of that. Remove the directory and the engine boots, types and tests unchanged.
Wire contract (client side: webapp `features/room/voice/`)
| Direction | Frame | |---|---| | client → engine | `{type:'session.context', version:1, room_id, room_name, surface:'room'\|'dm', capabilities?}` — after `session.config` and on every room change; a DM frame has `room_id: null` | | engine → client | `{type:'session.context.ack', version:1, room_id, bound, surface:'dm'\|'room'\|'refused', reason?}` | | engine → client | `{type:'ui.navigate', version:1, request_id, target:'dm'\|'room'\|'home', query?}` | | client → engine | `{type:'ui.navigated', version:1, request_id, ok, room_id?, room_name?, error?:'not_found'\|'ambiguous'\|'unsupported', candidates?}` |
What it does
- **`session-context.ts`** — parses and bounds the frame (Matrix room id grammar, name
cap, capability list) and builds the ack.
- **`room-binding.ts`** — the room a client names is a claim. It becomes the session's
origin only when the gateway bridge answers `state/voice-room-checks/<key>.request.json` with a verified `<key>.verdict.json` (owner and agent both joined). A different room releases the bound one *before* the wait; a newer frame supersedes a pending one; a refusal, a throw or silence leaves the session on the owner DM.
- **Origin handed to the core**: channel `ag2space`, target = the room id, header keys
`channel_kind: room` + `source_room_id`, the `room_context:` body line, the `[dm-only]` delivery note, and a `verify` that re-asks the gateway bridge when the result is delivered. Results therefore land as `results/proactive-result-<task>-<ts>.to-ag2space.txt` with `[channel: <room>]` first.
- **In-flight work** keeps the origin and recent context from when the `work` call
began (or the turn the relay agent's turn check sends), including across the relay agent's wait for the user to finish, its request check, a room switch or disconnect/reconnect. A pending transcription is omitted if the room changes, the turns provider is replaced, or the client disconnects; the new session's words never fill that old task. Re-announcing the room already bound, and a spoken farewell ("research X, bye"), change none of these, so the owner's words for that task are kept. Delivery still re-checks the original room's membership and honors `[dm-only]`.
- **Prompt**: a `ROOM:` context line while docked, and one system notice per actual
room change.
- **`current_room`** — "which room are you in?" is answered from the binding at the
moment it is asked, never from what the model remembers: the notice comes once per change, can be missed (a reconnect, a busy session) and can be outweighed by an earlier room still in context. It sits with `navigate_ui` in `voiceSurface()`, under the same condition, with its WHERE YOU ARE prompt rule and a turn-check rule. It is listed as an answering tool, so the turn never also goes to the core.
- **`navigate.ts` / `navigate-protocol.ts`** — the `navigate_ui` tool ("let's talk in my
DM", "take me to GTM in Investors", "go home"). It is contributed through `voiceSurface()`, so it is on the web voice session only, never the phone tool table, and only when this install has the gateway channel (`REMOTE_TASK_TOKEN` / `AG2_REMOTE_TOKEN` in the environment or in `channels/ag2space/.env` under the Claude config dir). Its NAVIGATION prompt rule is present under the same condition. The frame goes only to a client that announced the `ui.navigate` capability; any other attached client gets an immediate "update the app" answer.
Tests: `tests/ag2space-voice-*.test.ts` (skipped when this directory is absent).
Optionality check: move this directory aside, run `npx tsc --noEmit -p .` and `npm run test:ts` — both pass, with only the two `ag2space-voice-*` files reporting a skip.
Packaged desktop builds
A packaged engine runs on plain node and cannot import `tools.ts`. The desktop build bundles this skill to an `.mjs` the same way it bundles its private plugins and loads it through `SUTANDO_EXTERNAL_PLUGIN_DIRS`.
Read more
name: ag2space-voice description: "s.space voice surface — binds the room the s.space client announces to the voice session's origin, so work delegated by voice in a room is answered in that room, and moves the client by voice (navigate_ui)." when_to_use: "Loaded automatically at voice-agent startup as a manifest skill. Not slash-invoked. Only an s.space client sends the frames it handles; on any other install it stays idle."
s.space voice
Optional plugin for installs that talk to an s.space client (desktop app or webapp). The host engine carries a product-neutral *voice session origin* (`VoiceSessionOrigin`, `src/task-bridge.ts`) and offers client frames to skills (`setup(ctx)`, see [`skills/MANIFEST.md`](../MANIFEST.md)). This skill owns everything s.space specific on top of that. Remove the directory and the engine boots, types and tests unchanged.
Wire contract (client side: webapp `features/room/voice/`)
| Direction | Frame | |---|---| | client → engine | `{type:'session.context', version:1, room_id, room_name, surface:'room'\|'dm', capabilities?}` — after `session.config` and on every room change; a DM frame has `room_id: null` | | engine → client | `{type:'session.context.ack', version:1, room_id, bound, surface:'dm'\|'room'\|'refused', reason?}` | | engine → client | `{type:'ui.navigate', version:1, request_id, target:'dm'\|'room'\|'home', query?}` | | client → engine | `{type:'ui.navigated', version:1, request_id, ok, room_id?, room_name?, error?:'not_found'\|'ambiguous'\|'unsupported', candidates?}` |
What it does
- **`session-context.ts`** — parses and bounds the frame (Matrix room id grammar, name
cap, capability list) and builds the ack.
- **`room-binding.ts`** — the room a client names is a claim. It becomes the session's
origin only when the gateway bridge answers `state/voice-room-checks/<key>.request.json` with a verified `<key>.verdict.json` (owner and agent both joined). A different room releases the bound one *before* the wait; a newer frame supersedes a pending one; a refusal, a throw or silence leaves the session on the owner DM.
- **Origin handed to the core**: channel `ag2space`, target = the room id, header keys
`channel_kind: room` + `source_room_id`, the `room_context:` body line, the `[dm-only]` delivery note, and a `verify` that re-asks the gateway bridge when the result is delivered. Results therefore land as `results/proactive-result-<task>-<ts>.to-ag2space.txt` with `[channel: <room>]` first.
- **In-flight work** keeps the origin and recent context from when the `work` call
began (or the turn the relay agent's turn check sends), including across the relay agent's wait for the user to finish, its request check, a room switch or disconnect/reconnect. A pending transcription is omitted if the room changes, the turns provider is replaced, or the client disconnects; the new session's words never fill that old task. Re-announcing the room already bound, and a spoken farewell ("research X, bye"), change none of these, so the owner's words for that task are kept. Delivery still re-checks the original room's membership and honors `[dm-only]`.
- **Prompt**: a `ROOM:` context line while docked, and one system notice per actual
room change.
- **`current_room`** — "which room are you in?" is answered from the binding at the
moment it is asked, never from what the model remembers: the notice comes once per change, can be missed (a reconnect, a busy session) and can be outweighed by an earlier room still in context. It sits with `navigate_ui` in `voiceSurface()`, under the same condition, with its WHERE YOU ARE prompt rule and a turn-check rule. It is listed as an answering tool, so the turn never also goes to the core.
- **`navigate.ts` / `navigate-protocol.ts`** — the `navigate_ui` tool ("let's talk in my
DM", "take me to GTM in Investors", "go home"). It is contributed through `voiceSurface()`, so it is on the web voice session only, never the phone tool table, and only when this install has the gateway channel (`REMOTE_TASK_TOKEN` / `AG2_REMOTE_TOKEN` in the environment or in `channels/ag2space/.env` under the Claude config dir). Its NAVIGATION prompt rule is present under the same condition. The frame goes only to a client that announced the `ui.navigate` capability; any other attached client gets an immediate "update the app" answer.
Tests: `tests/ag2space-voice-*.test.ts` (skipped when this directory is absent).
Optionality check: move this directory aside, run `npx tsc --noEmit -p .` and `npm run test:ts` — both pass, with only the two `ag2space-voice-*` files reporting a skip.
Packaged desktop builds
A packaged engine runs on plain node and cannot import `tools.ts`. The desktop build bundles this skill to an `.mjs` the same way it bundles its private plugins and loads it through `SUTANDO_EXTERNAL_PLUGIN_DIRS`.
My AI Stand — Realtime by Day, Rewriting Itself by Night. Summon my AI superpower. Voice, vision, screen, meetings, calls when I'm engaged. Learns my patterns, ships its own code when I'm not. Runs across my Macs, interacts with people & their Stands.
Other skills on sutando.
More skills from other plugins.
browser-use
Direct browser control via CDP for web interaction: automation, scraping, testing, screenshots, and site/app work. Also covers…
agent-browser
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling…
Questions about ag2space-voice.
- Which coding agents can use ag2space-voice?
- ag2space-voice installs in Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and Windsurf with the skills CLI, using each agent's flag shown on this page. Flowy's auto-invocation, which fires a skill when your prompt matches, runs in Claude Code only.

