---
name: lead-finder
description: Find and qualify sales leads for a product or service with Monid under a strict user-provided spending cap, including verified business emails and evidence-based fit. Use when a user asks for prospecting, lead lists, decision-maker discovery, email enrichment, ICP-based account research, or optional LinkedIn, phone, and contact enrichment with explicit cost control.
---

# Lead Finder

Build a small, actionable lead list before expanding volume. Require the `monid` skill and Monid CLI; do not substitute web search, custom scraping, or another provider for the core retrieval and enrichment workflow.

At the start, verify that the `monid` skill, CLI, and an active API key are available. If any requirement is missing, stop retrieval, explain the missing dependency, and follow the current `monid` skill's setup instructions. Ask permission before installing software or saving a key. Never include Monid credentials in deliverables or skill files.

## Establish the brief

Collect or extract:

- the product or service, outcome, mechanism, and current proof;
- target industries, geography, company size, and buyer roles;
- observable buying triggers and disqualifiers;
- preferred lead count and required fields;
- an explicit maximum total spend in USD.

Read supplied websites, files, or notes before asking questions. Ask only for missing information that materially changes targeting or spend. Do not begin a paid Monid run until the spending cap is explicit.

State the working ICP, intended buyer experience, current targeting decision, riskiest assumption, and cheapest useful test. Recommend one segment when the brief is broad.

## Set the budget gate

Run `monid balance` and treat the user's cap as a hard limit for the complete workflow, not a target to consume.

Maintain a visible ledger with:

- maximum authorized spend;
- quoted worst-case cost before each run;
- reported cost after each run;
- cumulative nominal spend;
- remaining authorized budget.

Use the endpoint's advertised price and volume controls to calculate worst-case cost. Assume array limits may apply per query. Reserve enough budget for decision-maker and verified-business-email enrichment before buying broad account data.

Never start a run whose worst-case total would exceed the remaining authorization. If pricing is ambiguous, inspect again or ask the user to approve a clearly stated worst case. Do not interpret unused Monid balance as user authorization.

## Discover and inspect Monid tools

Follow the complete `monid` skill workflow:

1. Run focused `monid discover` queries for account search, people search, and work-email enrichment.
2. Inspect every candidate endpoint before running it.
3. Prefer verified endpoints, predictable pricing, free searches, and small first calls.
4. Read command hints and follow them when they improve the workflow.
5. Save completed results with the supported CLI output mechanism. If the installed CLI lacks the documented save flag, use its current supported retrieval behavior and avoid inventing flags.

Do not hard-code a provider, endpoint, schema, or price. Monid's live catalog and inspect output are authoritative.

## Find and qualify accounts

Search one narrow segment first. Start with 5–10 results unless a different volume is explicitly authorized.

Score each account using observable evidence:

- problem fit;
- buyer authority;
- operational complexity;
- buying trigger or growth signal;
- reachability;
- disqualifying evidence.

Reject keyword-only matches that do not fit the actual offer. Prefer a few strong accounts over a large generic list.

## Find decision-makers and emails

Identify the most relevant owner, founder, partner, executive, operations leader, or functional buyer at each shortlisted company. Use free people search before paid enrichment when available.

Provide a verified work email for every final lead whenever Monid can supply one within the authorized budget. Never guess, synthesize, or infer an email pattern. Label each address as verified, unverified, catch-all, or unavailable according to the source.

If the budget cannot support the requested number of verified emails, reduce the list and explain the tradeoff before spending more. If no verified email exists, provide the best public contact route and mark the lead incomplete; do not present it as delivery-ready.

## Deliver the first-pass list

Lead with a table containing:

| Priority | Person | Role | Company | Verified work email | LinkedIn or route | Evidence of fit | Recommended angle |
|---|---|---|---|---|---|---|---|

Then report:

- the working ICP and selection logic;
- rejected noise or important caveats;
- exact nominal cost and remaining authorization;
- a recommended outreach order;
- a stopping rule, such as testing the top three before buying more data.

Keep private raw responses and unnecessary personal data out of the deliverable.

## Pause before full analysis

After delivering emails, ask:

> Do you want a full analysis of the shortlisted people? I can verify their LinkedIn profiles, enrich their professional history, and create a personalized outreach checklist for each one. I will quote the additional cost before running anything paid.

Do not perform the full analysis merely because the user requested leads. Continue only after an explicit yes and a new or remaining budget sufficient for the proposed calls.

## Perform the optional LinkedIn analysis

For each approved person:

1. Search for and verify the LinkedIn profile by matching name, company, role, and location.
2. Use Monid to discover and inspect current profile, experience, about, posts, or enrichment endpoints.
3. Quote the worst-case cost and obtain authorization before paid calls.
4. Separate sourced facts from commercial inferences.
5. Produce a personalized `✅ Do / ❌ Don't` checklist grounded in specific roles, projects, numbers, posts, responsibilities, and business context.
6. Include one intelligent diagnostic question and one narrow pilot hypothesis.

Do not turn absence of public data into a personality judgment. Do not present growth, hiring, technology detection, or role history as proof of an internal problem.

## Offer optional enrichment

After the LinkedIn analysis, discover current Monid options and quote separate costs for useful add-ons, such as:

- business or mobile phone discovery;
- personal or secondary emails;
- recent posts and engagement;
- company technology, hiring, funding, or news signals;
- additional employees or buying-committee members.

Explain expected value, uncertainty, data type, and maximum incremental cost for each option. Never bundle optional enrichment into an earlier authorization. A phone result may be a mobile, company line, uncertain match, or unavailable; label it accurately.

## Protect people and the user

- Use professional contact data for legitimate B2B outreach.
- Minimize personal data and do not expose raw provider records.
- Follow applicable privacy, anti-spam, and telemarketing rules.
- Never claim consent, verification, or deliverability that the provider did not establish.
- Stop immediately when a Monid workspace control blocks a run and report how to resolve it.
- Clean up raw intermediate files after producing the approved deliverable unless the user asks to retain them.