I built a lead finder that refuses to guess

Description and use

Lead Finder plugin generates lead lists with sourced evidence, marking unknowns and applying strict scoring, improving lead reliability.

Problem

AI models produce confident but unverified company lists, making it hard to distinguish facts from guesses.

Outcome

The plugin produces a transparent, evidence-backed lead list, avoiding false positives and handling unknown data gracefully.

Workflow

• Install Lead Finder plugin via /plugin marketplace add machgtm/lead-finder
• Ask Claude to find 25 leads with specific criteria
• Claude gathers evidence and outputs JSON evidence files
• score.py processes evidence, applies rules, and ranks companies
• Output includes scorecard, evidence, and first question per company

Evidence

a fact needs a source, a guess becomes a question, and anything it couldn't check is "unknown".

Source

If coverage is below 50%, the fit is "unproven" and the company can't be marked "Call now".

Source

A company that clearly fails one (with a source) is skipped, whatever else it scores.

Source