Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Lead Extractor: Evidence-First Research

    1

    Give the agent a niche, geography, target count, and exclusions.

    $9

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Lead Extractor: Evidence-First Research

    Lead Extractor: Evidence-First Research

    Example session with this skill installed

    Find 20 independent commercial cleaning companies in Chicago. Verify each business, preserve separate branches, collect the full company schema, then use my connected browser to review LinkedIn People pages for visible owners, founders, operations leaders, and sales decision makers. Ask me to sign in if needed, stop for any security checkpoint, and return evidence-linked company and employee records ready for CSV or JSON.

    • Read your context and instructions
    • Compiled the lead extractor

    Research complete

    16 companies qualified, 3 need review, and 1 was excluded as a directory. The agent verified identity and service-area claims, preserved two distinct branches under one parent brand, and reviewed visible LinkedIn People-page results after the buyer completed sign-in.

    CompanyBranchWebsitePhoneEmail qualityStatusEvidence
    Lakeshore Facility CareChicago HQlakeshore.example(312) 555-0142Published generalQualifiedE1-E4
    Northstar Commercial CleanWest Loopnorthstar.example(312) 555-0188Direct publishedQualifiedE5-E8

    Visible decision makers

    EmployeeTitleRole groupSeniorityPriorityLinkedInEvidence
    Jordan LeeFounder & PresidentOwner / ExecutiveOwnerHighlinkedin.com/in/exampleL1
    Casey MorganDirector of OperationsOperationsDirectorHighlinkedin.com/in/example-2L2

    Evidence and safeguards

    • Company facts are tied to URLs, source types, observation times, and excerpts.
    • Employee fields reflect only information visible in the buyer-authorized browser session.
    • Parent-brand and operating-entity IDs keep separate locations from being merged incorrectly.
    • Unknown or conflicting values remain flagged for review; nothing is guessed.
    • The browser provider may retain the signed-in session, but the skill stores no password, MFA code, cookie, token, or browser profile.
    • No proxy rotation, CAPTCHA solving, stealth automation, CRM write, or outreach was performed.

    Deliverables

    The result includes the qualified shortlist, full company records, 19-field employee records, evidence ledger, duplicate decisions, diagnostics, limitations, and CSV/JSON-ready output. The three unresolved companies remain in a review queue with exact next steps.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    What this MCP skill is

    Lead Extractor turns a compatible AI agent connected through Agensi MCP into a structured lead-intelligence workbench. Instead of returning a loose list of company names, it guides the agent through the same evidence-first workflow as the Lead Extractor desktop workbench: discovery, company verification, branch-aware deduplication, contact research, LinkedIn employee review, scoring, evidence capture, checkpoints, and export-ready output.

    Give the agent a niche, geography, target count, and exclusions. It inspects the browser, search, page-reading, and MCP tools available to the buyer; explains what it can access; gathers candidates from diverse public sources; verifies each retained claim; and returns company and employee records using a fixed schema.

    What it can collect

    The skill covers the complete workbench field model, including:

    • Company identity, category, address, city, coordinates, and branch identity
    • Phones, emails, website, and published social profiles
    • Rating, review count, opening hours, and source identifiers
    • Discovery query, location, record type, and source strength
    • Evidence, observations, facts, assumptions, and quality flags
    • Qualification status, reason, lead score, and evidence score
    • Operator type and confidence, opportunity signals, and decision-maker fields
    • Contact quality, verification state, duplicate summary, and review status
    • Parent-brand and operating-entity IDs so separate branches are not flattened
    • LinkedIn employee name, visible headline, title, role group, seniority, priority, role tags, canonical profile URL, and evidence

    Missing values remain unknown rather than being invented.

    LinkedIn without unsafe credential handling

    When LinkedIn employee discovery is requested, the agent first checks whether the buyer's connected browser supports an authenticated session. If sign-in is needed, it opens LinkedIn and asks the buyer to sign in personally. The skill never asks for or stores passwords, MFA codes, cookies, session tokens, or browser-profile files.

    If the browser provider retains the session, the agent may reuse that authorized session. It then reviews a verified company's People page in bounded, sequential batches and extracts only visible, user-authorized fields. CAPTCHA, MFA, security checkpoints, 403/429 responses, and unusual-activity warnings always pause the workflow for manual action.

    Why it is better than generic lead-generation skills

    1. Fixed schema, not an improvised table — every run produces the same company and employee attributes, ready for JSON, CSV, or spreadsheet workflows.
    2. Evidence for every important claim — retained facts are tied to URLs, timestamps, source types, and visible excerpts.
    3. Branch-aware identity resolution — parent brands and operating locations stay separate, preventing contacts and evidence from leaking across branches.
    4. Authenticated LinkedIn support with human checkpoints — it can work through a buyer-controlled browser session without exposing credentials or bypassing platform protections.
    5. Resumable research — completed companies, evidence IDs, deduplication decisions, and unresolved checkpoints can be retained as non-secret workspace state.
    6. Conservative contact handling — direct, general, support-only, inferred, and unverified contacts remain clearly separated.
    7. Transparent limitations — inaccessible sources, conflicts, missing attributes, and tool failures stay visible instead of being filled with guesses.
    8. Agent-native and portable — no Windows app, creator-hosted service, or private project access is required. It works through the buyer's own compatible AI agent and connected tools.
    9. Buyer-controlled cost — optional Agensi data tools run only when requested and cost-approved.
    10. Safe by design — no proxy rotation, fingerprint spoofing, stealth plugins, CAPTCHA solving, cookie sharing, or rate-limit evasion.

    Typical workflow

    1. Buyer provides niche, geography, count, exclusions, and required attributes.
    2. Agent checks available tools and declares its coverage.
    3. Agent discovers and verifies companies from public sources.
    4. Agent resolves branches, contacts, and evidence.
    5. If requested, the buyer signs in to LinkedIn manually and authorizes a bounded People-page review.
    6. Agent classifies visible roles and decision makers.
    7. Agent returns a qualified shortlist, employee table, evidence ledger, diagnostics, limitations, and full CSV/JSON-ready records.

    The skill performs research and export only. It does not send outreach or write to a CRM unless the buyer makes a separate explicit request and connects an authorized tool.

    Live discovery depends on the web-search and browser tools available to the buyer's agent and on public-source accessibility. CAPTCHA, login, or paywall checkpoints require user action. The skill produces research and review output; it does not guarantee exhaustive coverage, email deliverability, or outreach execution.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 4 days ago

    Permissions

    Network Access
    Browser

    Browser and network access support live research. File access is optional for buyer-supplied lists or requested CSV/JSON exports. LinkedIn sign-in is completed manually by the buyer; no passwords, cookies, or session tokens are stored. No CRM connection or outreach delivery is required.

    Listed4 days ago

    What's inside

    Frequently Asked Questions

    No. The skill runs inside your compatible AI agent after you connect it through Agensi MCP. It does not require the creator's private application, a Windows installation, or a terminal.

    It gives your agent a repeatable lead-research method: define scope, discover candidates, verify identity, service and geography, preserve source-linked evidence, deduplicate branches, and return a review-ready shortlist.

    For live research, your agent needs web search or browser access. If those tools are unavailable, you can supply URLs, files, or excerpts for the skill to verify and organize.

    No. The core workflow uses the research tools already available to your agent. Optional Agensi data tools remain disabled unless you request them and approve any displayed cost.

    Every retained claim stays tied to a source URL and excerpt. Unknowns and conflicts remain visible, service areas are not treated as physical branches, and distinct branches stay separate unless strong identifiers support a merge.

    New in Sales & Lead Gen