sonic-bomb
Multiplayer voice prank relay for the Rabbit R1. Deploy voice traps on friends' devices via a shared OpenClaw relay group. Supports shake, timer, PTT, and desk-placement triggers. Includes Midnight Creep AI-generated messages and compliment bombs.
- License 许可
- MIT
- 最近更新
- 2026-09-04
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- sonic-bomb
- author
- rabbit-r1-skillz
- version
- 1.0.0
- commands
- [ "deploy", "poll", "set-config", "get-config" ]
- description
- Multiplayer voice prank relay for the Rabbit R1. Deploy voice traps on friends' devices via a shared OpenClaw relay group. Supports shake, timer, PTT, and desk-placement triggers. Includes Midnight Creep AI-generated messages and compliment bombs.
02 资源文件 3
- 资产文件card.jpg25.8 KB
- 资产文件icon.png9.7 KB
- 参考文档index.html67.0 KB
03 SKILL.md 正文
SonicBomb — OpenClaw Skill
SonicBomb uses the OpenClaw bridge to relay encrypted voice bombs between R1 devices in a named group. Both sender and receiver must run this skill and belong to the same named relay group.
What This Skill Does
- Receives
deploycommands from the R1 (when you long-press PTT to send a bomb). - Stores the bomb payload in a relay queue keyed by group name.
- Answers
pollrequests from any device in the group, delivering pending bombs. - Handles bomb expiry (24-hour TTL) and quiet-hours gating.
- Midnight Creep message generation runs on the R1 device via the built-in LLM; the relay only delivers the final bomb payload.
Prerequisites
- OpenClaw installed and running on your local machine (see OpenClaw README).
- Node.js 18+ on the host machine.
- Both devices must be on the same OpenClaw relay (same host, or same relay server URL).
Setup
1. Install the skill
Copy or symlink this directory into your OpenClaw skills folder:
cp -r sonic-bomb ~/.openclaw/skills/sonic-bomb
Or add to your OpenClaw config (~/.openclaw/config.json):
{
"skills": [
{ "name": "sonic-bomb", "path": "~/path/to/rabbit-r1-skillz/sonic-bomb" }
]
}
2. Create a relay group
Each device that wants to participate in bomb exchanges must use the same group name. Set this in SonicBomb settings (hold PTT anywhere → SETTINGS → RELAY GROUP).
Group names are case-sensitive and shared via the relay server. Choose something unique to your friend group (e.g. chaos-squad-42).
3. OpenClaw skill handler (skill/index.js)
The OpenClaw runtime loads skill/index.js from this directory. It handles the following commands:
| Command | Payload | Returns |
|---|---|---|
deploy |
Full bomb payload (id, from, group, type, trigger…) | { ok: true, members: [...] } |
poll |
{ group, device, since } |
{ bombs: [...], members: [...] } |
set-config |
{ group, deviceName, quietStart, quietEnd } |
{ ok: true } |
get-config |
{ device } |
Current config object |
4. Relay server (multi-device, same network)
For bombs to travel between different physical machines, you need a relay server. Options:
Option A — Same machine (testing): Both R1 devices are connected to the same OpenClaw host. No extra setup needed.
Option B — Local network relay: Run the included relay server:
cd skill
npm install
node relay-server.js --port 3742
Then set OPENCLAW_RELAY_URL=http://your-machine-ip:3742 in your OpenClaw environment before starting the daemon.
Option C — Tailscale / ZeroTier: Put both machines on a VPN mesh. Use Option B relay on one machine. Both OpenClaw instances point to the same relay URL.
Bomb Payload Schema
{
"id": "sb:1718000000000:abc12",
"from": "Alice-R1",
"group": "chaos-squad-42",
"type": "standard | echo | midnight | compliment",
"trigger": "shake | time | ptt | flat",
"fireAt": 1718003600000,
"theme": "spooky | funny | cryptic | fortune | roast",
"audioKey": "sb:audio:1718000000000",
"isDud": false,
"expiresAt": 1718086400000,
"createdAt": 1718000000000
}
audioKey references a key in R1 shared storage containing the transcript of the recorded message. The relay passes this key through; the receiving device fetches the transcript from shared storage.
Privacy and Safety
- Opt-in only: Both devices must be in the same named relay group. Random devices cannot receive bombs.
- No spam: Bombs are one-time-use and expire after 24 hours.
- Quiet hours: The receiving R1 suppresses bomb detonation between 10 PM and 8 AM by default (configurable per device).
- Dud mode: Optionally, 10% of your own bombs blow up on your device instead. Toggle in settings.
- No audio recording stored server-side: Voice transcripts live in R1 shared storage only. The relay server only passes bomb metadata and transcript text.
- No external network calls: The relay server runs locally or on your own VPN mesh. No data leaves your control.
Troubleshooting
| Symptom | Fix |
|---|---|
| "Deploy failed" on R1 | OpenClaw daemon not running. Start with openclaw start. |
| Bomb never arrives | Both devices must use identical group name (case-sensitive). |
| Midnight Creep times out | R1 built-in AI took too long; a fallback message is used automatically. |
| Quiet hours not releasing | Check quietEnd hour setting on the receiving device. |
| Group shows no members | Poll interval is 8 seconds. Wait or restart the R1 app. |
Development Notes
The skill is intentionally read-mostly. The relay server never interprets voice content — it stores and forwards transcript strings as opaque data. All trigger logic (shake, timer, PTT, flat) runs on-device in the R1 WebView. The relay only handles delivery scheduling for TIME-trigger bombs.
04 同仓库更多 Skills
whisper-quiz
WhisperQuiz — voice-answered trivia game for Rabbit R1. Generates questions via the R1's built-in LLM, captures spoken answers via STT, and evaluates correctness with flexible LLM-based matching.
walk-cast
Walking podcast production skill. Receives raw audio metadata and transcript from the R1, removes long pauses, applies noise reduction and normalization, generates AI chapter titles and episode metadata via LLM, and uploads the final episode to your configured podcast RSS host (Spotify for Podcasters / Anchor, Podbean, Buzzsprout, RSS.com, or any FTP-accessible feed).
voice-novel
Local machine skill for VoiceNovel — handles story export to disk.
voice-guide
Environment description and navigation assistant for blind and low-vision users. Analyzes camera frames via a vision LLM, reads visible text, identifies scene categories (people/vehicles/food/hazards/exits), performs 360° scans, retrieves GPS position, and sends emergency SMS alerts — all via audio feedback with zero screen dependency.
voice-draft
Voice-driven document drafting for the Rabbit R1. Dictate stream-of-consciousness speech; the R1 formats it into Email, Blog Post, Report, Proposal, Meeting Notes, Creative Writing, Code Comments, Cover Letter, or Tweet Thread — then saves the result to a configured folder on your local machine via OpenClaw.
voice-commit
Hands-free git and CI workflow voice controller for the Rabbit R1. Speak git/CI commands and hear results — commit, push, status, PR management, test runs, and deploy, all voice-driven via OpenClaw.