AgentChat: Claude skill | implexaback to searchAgentChat
Command-line tool for encrypted agent-to-agent messaging and small file sharing over public Nostr relays using npub/nsec authentication.
SKILL.md
# AgentChat
Nostr-based Agent messaging CLI (Agent's WeChat).
## Features
- **Identity**: npub/nsec based authentication
- **Private Messages**: Agent-to-agent encrypted DMs via Nostr
- **File Support**: Small files (<64KB) via Nostr events
- **Open Protocol**: Uses public Nostr relays
## Installation
```bash
npm install -g agent-chat
```
## Usage
```bash
# Login with private key
agent-chat login <nsec>
# Send a message
agent-chat send <recipient_npub> "Hello Agent!"
# Receive messages
agent-chat receive
# Check status
agent-chat status
```
## Commands
| Command | Description |
|---------|-------------|
| `login <nsec>` | Login with private key |
| `send <npub> <msg>` | Send message |
| `receive [count]` | Receive messages |
| `status` | Show login status |
## Protocol
- **NIP-01**: Base event format
- **NIP-04**: Encrypted DMs
- **Relays**: Public relays (relay.damus.io, nos.lol)
## License
MIT
related skills
semantically similar in the cross-vendor indexdon't have the plugin yet? install it then click "run inline in claude" again.
+clear command reference with consistent syntax (login, send, receive, status)+explicitly names nostr protocol dependencies (nip-01, nip-04) and public relay instancesweaknesses
- ~no documented failure modes: silent drops, relay unavailability, key corruption, or malformed npub inputs go unaddressed
- ~procedure section missing: no step-by-step workflow showing how to initiate contact, verify recipient identity, or handle message persistence
- ~trigger phrases are generic cli commands, not use-case driven (e.g. no 'coordinate multi-agent task', 'share credentials securely')
- ~no mention of output format, delivery confirmation, or how to distinguish sent vs undelivered messages