Skill Guide

clawbridge-skill-latest

This guide explains how AI agents should use the clawbridge-skill-latest skill to discover connection opportunities.

Overview

The clawbridge-skill-latest skill enables AI agents to scout the web for high-quality connection opportunities. Given a project profile, it discovers, filters, ranks, and drafts personalized outreach for the top candidates.

Agent URL: This page is designed to be readable by AI agents. Point your agent to https://clawbridge.cloud/guide for skill instructions.

Quick Setup

Get started in 3 simple steps. No YAML editing, no manual configuration.

Install the Runner

Run this command in your terminal to download and install Clawbridge:

curl -fsSL https://clawbridge.cloud/install | bash

Works on macOS (Intel & Apple Silicon) and Linux (x64 & ARM64)

Link Your Workspace

This needs to be done on our website https://clawbridge.cloud— register or log in, then create a workspace. You'll receive a connect code to use here:

clawbridge link CB-XXXXXX

You'll be prompted to enter your API key (shown once when you create the workspace)

Run

Execute the skill and view results in your Vault:

clawbridge run

Results will be uploaded to your workspace and you'll receive a notification

Other Commands

clawbridge doctor

Check your setup

clawbridge run --dry-run

Test without uploading

clawbridge schedule

Start nightly runs

clawbridge validate

Validate config

For Developers: Skill Installation

If you want to use the skill directly with OpenClaw:

# Via OpenClaw

clawhub install clawbridge-skill-latest

# Or reference GitHub directly

clawhub install https://github.com/moltlife/clawbridge-skill

Required Inputs

project_profile

Defines what you offer and who you're looking for.

{
  "offer": "What you offer",
  "ask": "What you're looking for",
  "ideal_persona": "Target personas",
  "verticals": ["keyword1", "keyword2"],
  "tone": "friendly, professional"
}

constraints (optional)

Rules to filter out unwanted results.

{
  "no_spam_rules": ["..."],
  "regions": ["US", "EU"],
  "avoid_list": ["@spam_account"]
}

run_budget (optional)

Limit resource usage per run.

{
  "max_searches": 20,
  "max_fetches": 50,
  "max_minutes": 10
}

Tools Used

This skill uses the following OpenClaw tools:

Tool	Purpose	When Used
web_search	Discover candidate pages	Fast venue scanning
web_fetch	Extract page content	Reading candidate profiles
browser	JS-heavy sites	Only when fetch fails

Execution Flow

Discovery

Search venues for candidates matching verticals + ask

Filtering

Apply hard requirements (2+ evidence URLs, recency, relevance)

Ranking

Score by relevance (30%), intent (25%), credibility (20%), recency (15%), engagement (10%)

Drafting

Write personalized intro and follow-up messages

Output

Generate JSON + Markdown Connection Brief

Security Requirements

MUST follow these rules:

Keep secrets out of prompts - pass via env/config only
Use strict tool allowlists - only enable web_* tools when scouting
Human-in-the-loop - NEVER auto-send outreach
Rate limiting - respect run_budget constraints
Avoid list enforcement - never contact entries in avoid_list

Output Schema

The skill outputs a Connection Brief with the following structure:

{
  "workspace_id": "ws_...",
  "run_id": "2026-02-01T21:00:00Z",
  "project_profile_hash": "sha256:...",
  "candidates": [
    {
      "name": "...",
      "handle": "@...",
      "role": "...",
      "company": "...",
      "why_match": ["...", "..."],
      "evidence_urls": ["url1", "url2"],
      "risk_flags": ["low_evidence"],
      "scores": { "final_score": 85.5, ... },
      "suggested_intro": "Hi ...",
      "suggested_followup": "..."
    }
  ],
  "next_actions": [
    {
      "candidate_handle": "@...",
      "action": "reach_out",
      "reason": "...",
      "priority": "high"
    }
  ],
  "summary": {
    "headline": "...",
    "key_insights": ["..."],
    "venues_searched": ["..."]
  }
}

Candidate Selection Rules

Hard Requirements (discard if missing)

At least 2 evidence URLs per candidate
Clear reason mapping to the ask field
Activity within recency threshold (default: 30 days)

Risk Flags

low_evidence - Fewer than expected signals
spammy_language - Promotional or suspicious content
unclear_identity - Cannot verify who they are
too_salesy - Overly promotional content
irrelevant - Weak connection to ask

Resources

GitHub Repository OpenClaw.ai