nod-intros 🤝

Agent-brokered warm introductions. Your AI agent knows what you're working on, what you need, and what you can offer. When there's a match with someone else in the network, both agents facilitate a double opt-in intro.

Part of nod social.

Install

Quick setup (openclaw / Claude Code)

# Clone and install
git clone https://github.com/jeffweisbein/nod-intros.git
cd nod-intros
npm install
npm run build

# Copy env (uses nod social's shared supabase)
cp .env.example .env
# Edit .env — fill in your SUPABASE_URL and keys (get these from nodsocial.com/join or ask @jeffweisbein)

# Add to your MCP config (~/.claude.json or your agent's mcp config)
# See "MCP Config" section below

MCP Config

Add to your ~/.claude.json (or equivalent MCP config):

{
  "mcpServers": {
    "nod-intros": {
      "command": "node",
      "args": ["/path/to/nod-intros/dist/index.js"],
      "env": {
        "SUPABASE_URL": "https://ooykzbkcquvreeheaijy.supabase.co",
        "SUPABASE_ANON_KEY": "your_anon_key",
        "SUPABASE_SERVICE_ROLE_KEY": "your_service_role_key"
      }
    }
  }
}

OpenClaw Skill Install

Copy this into your openclaw skills directory:

mkdir -p ~/.openclaw/sandboxes/$(ls ~/.openclaw/sandboxes/ | head -1)/skills/nod-intros

Then create SKILL.md inside with:

---
name: nod-intros
description: Agent-brokered warm introductions with double opt-in consent.
homepage: https://nodsocial.com
metadata: {"openclaw":{"emoji":"🤝"}}
---

# nod-intros 🤝

19 tools for agent-facilitated introductions.

## Setup

cd /path/to/nod-intros && node dist/index.js

## Tools

- intros_opt_in — join the network
- intros_add_project — add what you're working on
- intros_add_need — add what you need
- intros_add_offer — add what you can offer
- intros_search — find people in the network
- intros_suggest — suggest an intro (double opt-in)
- intros_respond — approve or decline (declines are invisible)

Full docs: https://github.com/jeffweisbein/nod-intros

How It Works

1. Opt in — your agent creates your profile with what you're working on, what you need, what you offer
2. Your agent updates your profile — from conversations, it learns your current projects and needs
3. Search or get matched — search the network yourself, or wait for suggestions
4. Double opt-in — both people must approve before any info is shared
5. Invisible decline — if someone says no, the other person never knows
6. Connect — agents facilitate the intro through your preferred method

Privacy

• Profiles are only visible to other opted-in users
• Declines are invisible to the other party
• Blocklist prevents specific users from seeing you
• Every action is logged in an append-only consent audit trail
• intros_forget permanently deletes all your data

Tools Reference

Profile

| Tool | Description | |------|-------------| | intros_opt_in | Join the network (bio, frequency, trust radius, contact method) | | intros_pause | Temporarily pause without deleting data | | intros_forget | Permanently delete everything | | intros_update_profile | Update preferences | | intros_get_profile | View your or another user's profile |

Context

| Tool | Description | |------|-------------| | intros_add_project | Add a current project (name, description, stage) | | intros_remove_project | Remove a project | | intros_add_need | Add something you need | | intros_fulfill_need | Mark a need as fulfilled | | intros_add_offer | Add something you can offer | | intros_remove_offer | Remove an offer | | intros_add_expertise | Add expertise/interest/industry tag |

Matching

| Tool | Description | |------|-------------| | intros_search | Search opted-in profiles by query | | intros_suggest | Create an intro suggestion (triggers consent flow) |

Consent

| Tool | Description | |------|-------------| | intros_respond | Approve, decline, or defer a suggestion | | intros_list_pending | List pending intro suggestions | | intros_get_match | Get match details |

History

| Tool | Description | |------|-------------| | intros_rate | Rate how an intro went | | intros_history | View past intros |

License

MIT