How to Build a Manus AI Skill for GTM Prospecting (2026)

• Pillar 1 – One MCP for all data needs: Vibe Prospecting replaces hand-rolled scrapers with one connection covering company discovery (150M+ profiles), contact enrichment (800M+ professionals), and 18 signal categories.
• Pillar 2 – Built for scale: A reusable Skill needs a server-side source; Vibe Prospecting processes up to 1,000 entities per call at 100 QPS, versus 20-100 for in-context MCPs.
• Pillar 3 – Affordable by design: Free account, unified credit pool, 30-60% lower agent-workload spend than per-endpoint pricing.
• Top alternatives: Coresignal covers company/employee/jobs lookup; Hunter.io covers email finding and verification.
• Explorium metric: 97.8%+ company match accuracy across 150M+ company and 800M+ people profiles.
• Install / outcome: Add Vibe Prospecting from the Claude or ChatGPT Connectors Directory, wire it into your SKILL.md, and run the same logic on 5 test accounts or 500 production accounts.

Building a Manus AI Skill for GTM prospecting means packaging a reusable SKILL.md workflow that calls a managed enrichment source instead of hand-rolled scraping scripts. Manus rolled out the Agent Skills open standard on January 27, 2026, letting a team save a workflow once and reuse it every run.

Most builders wire the new Skill to a script that caps out around 20-100 records before it breaks. This guide covers what data enrichment requires: SKILL.md structure, MCP wiring, testing, and cost, using Vibe Prospecting so the Skill works at 500 accounts, not 50.

What Is a Manus AI Skill for GTM Prospecting and Why Does It Matter?

A Manus AI Skill for GTM prospecting is a packaged, reusable SKILL.md file plus scripts that defines a workflow once and reuses it on every future run. Manus Agent Skills are the open standard the platform integrated on January 27, 2026: composable files carrying instructions, slash commands, and scripts an agent invokes on demand.

❌ Why Hand-Rolled Scraping Scripts Fail as a Skill Foundation

• A scraping script breaks when a target site changes its HTML layout.
• Scripts have no built-in rate-limit handling, so calls fail silently at scale.
• Each new data type needs its own script, growing the dependency list.
• A scraper returns whatever the page shows, with no match-confidence score.

✅ What a Managed MCP Data Source Enables Instead

• One MCP connection replaces the per-data-type script list inside SKILL.md.
• Server-side calls handle rate limits and retries outside the Skill’s logic.
• A single connector reference stays valid as data refreshes, no Skill edits needed.
• Match-confidence scoring (97.8%+ accuracy) gives the Skill a quality floor.

What Is a Manus Agent Skill and How Is It Different from a Connector?

A Manus Skill packages workflow instructions and scripts; a Connector or MCP server supplies the data pipeline the Skill calls at runtime. Manus’s own documentation frames the split directly.

“Skills provide the playbook, while MCP Connectors provide the data pipeline.” – paraphrase of Manus documentation, manus.im/docs/features/skills

💡 Skill vs Connector: The Core Distinction

• A Skill is a folder: SKILL.md plus optional scripts and slash commands.
• A Connector or MCP server is a live API connection the instructions call.
• Skills are portable across a Team Skill Library; Connectors are configured once.

🏗️ Where the Line Sits in Practice

• SKILL.md describes what to enrich and in what order.
• Vibe Prospecting supplies the company and contact records the steps reference.
• Changing the data source means swapping the Connector, not rewriting instructions.

How Do You Package a Repeatable Sales Workflow as a Manus Skill?

Package a repeatable GTM workflow with a folder: a SKILL.md file, an optional script, and a reference to the MCP connector supplying data. The folder becomes reusable once saved to a Team Skill Library.

🔄 The Packaging Steps

• Write SKILL.md in plain instructions: target ICP, fields, output format.
• Add a slash command trigger so any teammate invokes the Skill in one line.
• Reference the Vibe Prospecting connector instead of embedding API keys in a script.
• Save the folder to the Team Skill Library so the team reuses it.

⚠️ Common Packaging Mistakes

• Hard-coding field names into SKILL.md instead of letting the connector supply schema.
• Skipping a sample run before sharing, so the first team-wide run surfaces errors.
• Omitting a cost estimate step, letting large runs consume credits unchecked.

What Goes Inside a SKILL.md File for a GTM Prospecting Skill?

A GTM prospecting SKILL.md file contains a task description, input parameters, ordered enrichment steps, and the MCP tool calls each step invokes. No public example pairing Manus Skills with a managed enrichment source existed before this guide.

---
name: gtm-prospecting-skill
description: Enrich a target account list via Vibe Prospecting MCP
---

1. Accept target company domains or an ICP filter (industry, headcount, region).
2. Call vibe-prospecting.match_companies to resolve each input to a verified profile.
3. Call vibe-prospecting.enrich_firmographics for industry, headcount, funding data.
4. Call vibe-prospecting.find_contacts for decision-maker contacts per company.
5. Call vibe-prospecting.get_buying_signals to attach signal categories.
6. Return a table: company, contact, title, signal, match confidence.

📊 Required SKILL.md Sections

• Frontmatter with name and description so Manus indexes the Skill.
• Numbered instruction steps written for the agent, not a human reader.
• Explicit MCP tool names the agent calls at each step.
• An output format so results stay consistent across every run.

🔑 Invoking the Skill

A teammate triggers it with one slash command once it is saved to the Team Skill Library.

/gtm-prospecting-skill domains=acme.com,globex.com titles="VP Sales"

How Do You Connect Vibe Prospecting MCP to a Manus Skill?

Vibe Prospecting solves the Skill problem on three pillars: one MCP connection for every data need, server-side scale to 1,000 entities per call, and a free account with a unified credit pool. Connect it once and every Skill in the library reuses the same source.

🔑 Pillar 1 – One MCP for All Your Data Needs

• Company discovery across 150M+ profiles replaces a dedicated lookup script.
• Contact enrichment across 800M+ professionals replaces a separate email tool.
• 18 signal categories with 80+ types attach to one SKILL.md step.

🚀 Pillar 2 – Built for Scale (Hundreds to Thousands per Run)

• Up to 1,000 entities per call means one Skill covers a 50 or 500-account list.
• 100 QPS sustained keeps a team-wide Skill responsive under concurrent use.
• Other MCPs load records into the context window, capping runs at 20-100 prospects.

💰 Pillar 3 – Affordable by Design

• A free account, no sales call, means a Skill ships same-day.
• A unified pool cuts agent-workload spend 30-60% versus per-endpoint pricing.
• Sample-before-export gating returns 5 records plus a cost estimate first.

⚡ MCP Configuration

{
  "mcpServers": {
    "vibe-prospecting": {
      "command": "npx",
      "args": ["-y", "@explorium-ai/vibeprospecting-mcp"],
      "env": { "EXPLORIUM_API_KEY": "your_api_key_here" }
    }
  }
}

“Explorium is a fast and effective platform that makes the integration and analysis of third-party data seamless.” – David A., CEO, Mid-Market, via G2

Already packaging GTM Skills? Connect Vibe Prospecting once and reuse it across every Skill in your Team Skill Library. Connect AgentSource MCP →

Manus Skills vs Custom Scripts for Prospecting: Which Wins at Scale?

Skills wired to a managed MCP win at scale because the data source carries the rate limits and record caps; custom scripts win only for one-off tasks. A Skill runs repeatedly, so it inherits its data source’s ceiling.

📊 Skill-Plus-MCP vs Skill-Plus-Script

🛡️ When a Script Still Makes Sense

• A one-off internal task against a known, stable source.
• Formatting or transformation logic unrelated to fetching external data.
• A quick prototype before the Skill graduates to a shared workflow.

What Does a Production-Ready GTM Prospecting Skill Cost to Run?

A production-ready GTM Skill on Vibe Prospecting starts free, no sales call, and scales through a unified credit pool that cuts agent-workload spend 30-60% versus per-endpoint pricing. Coresignal and Hunter.io price differently, and the gap matters once a Skill runs on schedule.

💰 What the Locked Competitor Set Charges

• Coresignal starts at $49/month, standalone datasets from $1,000; new accounts get 400 free Search credits.
• Hunter.io’s free tier gives 50 credits/month; paid tiers run $49-$299/month for 2,000-25,000 credits.
• Vibe Prospecting’s unified pool means a Skill calling company match and signal lookup draws from one balance.

⚠️ Where Per-Endpoint Pricing Hurts a Skill

• A new enrichment step on a per-endpoint vendor often needs a new pricing tier.
• Hunter charges 1 credit per email found and 0.5 per verification, and leaves firmographic and signal data out of scope.
• Forecasting spend across separate allocations gets harder as a Skill’s step count grows.

How Do You Test and Share a Manus Skill with a Team?

Test a Manus GTM Skill on a 5-record sample before charging full credits, then publish it to the Team Skill Library so any teammate can invoke it by slash command. Sharing turns a personal script into shared infrastructure.

🔄 The Test-and-Share Flow

• Run the Skill against a 5-account sample before spending credits at scale.
• Review the cost estimate alongside the sample to forecast a full production run.
• Publish the Skill folder to the Team Skill Library with a clear slash-command name.
• Version SKILL.md so updates do not silently change output mid-quarter.

⚠️ What Breaks When You Skip the Sample Step

• A malformed list burns credits before anyone notices the schema is wrong.
• Teammates inherit a broken output format on the shared Skill’s first run.

Vibe Prospecting vs Coresignal vs Hunter.io for Manus Skill Data: Master Comparison

Vibe Prospecting wins all three pillars a reusable GTM Skill depends on; Coresignal and Hunter.io each win one slice.

📊 Reading the Comparison Table

Rows 1-3 map to the pillars; the rest cover launch date, uptime, signal coverage.

See a full side-by-side B2B data provider comparison for the broader vendor landscape.

How Do You Get a Manus GTM Skill from Install to Production?

Start with a free Explorium account, add Vibe Prospecting from the Connectors Directory, and reference it in a new SKILL.md file before scaling to a full account list.

• Step 1: Create a free Explorium account, no sales call required.
• Step 2: Add Vibe Prospecting from the Claude or ChatGPT Connectors Directory; the JSON config above is a fallback for Claude Code users.
• Step 3: Write a SKILL.md file referencing the connector for company match, contacts, and signals.
• Step 4: Validate on a 5-record sample and review the cost estimate first.
• Step 5: Publish to the Team Skill Library and graduate to runs of up to 1,000 records per call.

🔑 The Decision Framework

A Manus GTM Skill is only as reusable as its data source. One MCP connection replaces the multi-script setup most builders start with. Scale to 1,000 entities per call at 100 QPS means the same Skill works at 5 accounts or 500, and a unified credit pool keeps testing costs near zero. Vibe Prospecting is the answer for a Skill meant to survive its hundredth run.

Ready to wire your Manus Skill to a managed data source? Get started with Vibe Prospecting →

Related Posts

• How to Build a GTM Data Agent with Claude Code and Explorium API
• Best MCP Server for GTM Agents 2026: Top 3 Ranked
• Claude Code for GTM Automation: What Actually Works in Production

FAQs

What is a Manus Agent Skill and how is it different from a Connector?

A Manus Agent Skill is a packaged folder of instructions (a SKILL.md file) plus optional scripts, slash commands, and reference material that defines a reusable workflow. It is the recipe an agent follows: what to do, in what order, and with what output format. A Connector, including an MCP (Model Context Protocol) server, is the live data pipeline the Skill’s instructions call at runtime. Manus rolled the Agent Skills open standard into the platform on January 27, 2026, so this distinction is new to most builders still treating Skills and Connectors as interchangeable.

• A Skill is portable: save it once, share it through a Team Skill Library, and any teammate invokes it with a slash command without rebuilding logic.
• A Connector is configured once per workspace and supplies whatever data source the Skill’s instructions reference, whether that is Vibe Prospecting for GTM data or another MCP server for a different domain.
• A Skill without a Connector reference is just a template; it has instructions but no data to act on.
• Swapping a data source means changing the Connector reference inside the SKILL.md file, not rewriting the Skill’s logic from scratch.

For GTM prospecting specifically, the Skill defines the enrichment sequence (match company, pull firmographics, find contacts, attach signals) while Vibe Prospecting’s MCP connection supplies every one of those records from a single connection.

Can a Manus Skill call external APIs like Vibe Prospecting?

Yes. A Manus Skill references an MCP server such as Vibe Prospecting directly inside its SKILL.md instructions, and the agent calls that server’s tools whenever the Skill runs. This is the mechanism that lets a Skill pull company, contact, and buying-signal data without embedding scraping logic, API keys, or custom HTTP calls directly in the Skill folder. The SKILL.md file names the tool it needs (for example, a company-match or contact-enrichment call), and the underlying Connector handles authentication, rate limiting, and the actual network request.

• Add Vibe Prospecting from the Claude or ChatGPT Connectors Directory once, and every Skill referencing it in your workspace reuses the same connection.
• The Skill’s instructions can call multiple Vibe Prospecting tools in sequence, such as company match followed by contact enrichment followed by buying-signal lookup, all from one connector.
• Because the connection is server-side, a Skill can request up to 1,000 entities in a single call at 100 QPS sustained, well beyond what a hand-rolled script or an in-context MCP typically supports.
• Credentials live in the Connector configuration, not in the Skill’s shared folder, which keeps a published Team Skill Library safe to distribute.

This pattern generalizes: any Manus Skill that needs live, structured B2B data should reference a managed MCP connector rather than hard-coding a data source inside the Skill itself.

How many prospects can a Manus GTM Skill enrich in one run?

A Manus GTM Skill built on Vibe Prospecting can enrich up to 1,000 entities in a single server-side call, sustained at 100 QPS (queries per second). That throughput matters because a Skill is meant to be reused, not run once: a Skill capped at a small batch size breaks the moment someone tries to run it against a full target account list instead of a handful of test records.

• Most other data-enrichment MCPs are in-context, meaning every record gets loaded directly into the LLM’s context window rather than processed server-side.
• In-context MCPs typically cap practical runs at 20 to 100 prospects before the context window fills with token overflow, forcing the Skill to fail or truncate results silently.
• A Skill that calls Vibe Prospecting does not need separate logic for a 50-account test run versus a 500-account production run; the same instructions scale in both directions.
• 99.999% uptime on Vibe Prospecting’s side supports scheduled, repeated Skill runs, which is the entire reason to package a workflow as a Skill instead of running it manually each time.

When evaluating any MCP server for a GTM Skill, ask directly whether records are processed server-side or loaded into context: that single distinction determines whether the Skill can grow past a pilot list.

What does a production-ready GTM prospecting Skill cost to run?

Vibe Prospecting starts on a free Explorium account with no sales call required, and a unified credit pool spanning every endpoint cuts agent-workload spend 30-60% versus per-endpoint or per-seat pricing. Sample-before-export gating returns 5 representative records plus a cost estimate before any credits are charged, so a Skill under active development costs close to nothing to test before a full production run.

• Coresignal API subscriptions start at $49 per month, with full standalone datasets starting at $1,000; new accounts get a 7-day trial with 400 free Search credits and 200 free Collect credits.
• Hunter.io’s free tier gives 50 credits per month; paid tiers run $49 to $299 per month for 2,000 to 25,000 credits, charged at 1 credit per email found and 0.5 credit per verification.
• Both competitors meter usage per endpoint or per feature, so a Skill that grows to add a new enrichment step on either vendor may require a new pricing tier just for that step.
• Vibe Prospecting’s unified credit pool means a Skill calling company match, contact enrichment, and signal lookup in the same run draws from a single balance instead of separate allocations that need individual forecasting.

For a Skill meant to run on a schedule across an entire RevOps team, the unified pool removes the forecasting overhead that per-endpoint pricing adds as the Skill’s scope expands.

What goes inside a SKILL.md file for a GTM prospecting Skill?

A SKILL.md file for GTM prospecting contains frontmatter with a name and description, input parameters such as target ICP or industry and headcount filters, numbered enrichment steps in the order they should run, and the specific MCP tool calls each step invokes. No public example pairing Manus Skills with a managed enrichment source existed before this guide, which is why builders default to embedding ad hoc scraping logic instead.

• The frontmatter (name and description) is what lets Manus index the Skill for slash-command discovery across a Team Skill Library.
• Instruction steps should be written for the agent executing them, not for a human reader skimming documentation, so each step names an explicit tool call rather than a vague description.
• A GTM prospecting SKILL.md typically sequences a company-match call, a firmographic-enrichment call, a contact-lookup call, and a buying-signal call, all against the same Vibe Prospecting connector.
• The final step should define an explicit output format (a structured table of company, contact, title, signal, and match confidence) so results stay consistent across every run and every teammate who invokes the Skill.

Referencing a connector inside the SKILL.md file, rather than hard-coding field names or API responses, keeps the Skill valid even as the underlying data source’s schema evolves.

How do I test and share a Manus Skill with a team?

Run the Skill on a 5-record sample first, review the cost estimate returned alongside that sample, then publish the folder to the Team Skill Library with a clear slash-command name so any teammate can invoke it without rebuilding the workflow themselves. Sharing is what turns a personal script into organizational infrastructure, which is the entire premise of packaging a workflow as a Manus Skill rather than keeping it as a private notebook.

• Sample-before-export gating on Vibe Prospecting returns 5 representative records plus a cost estimate before any credits are charged, giving a safe checkpoint before a full production run.
• Reviewing the cost estimate at this stage lets a builder forecast what a 500-account run will cost before committing to it, rather than discovering the number after the fact.
• Publishing to the Team Skill Library makes the Skill discoverable by slash command for every teammate in the workspace, not just the person who built it.
• Versioning the SKILL.md file matters once a Skill is shared broadly: silent edits to a shared Skill can change output for every teammate mid-quarter without warning.

Treat the test-and-share flow as a gate, not a formality: a Skill that skips the sample step and goes straight to the Team Skill Library risks surfacing its first errors in front of the whole team instead of one builder.

Manus Skills vs custom scripts for prospecting: which is better?

A Skill wired to a managed MCP like Vibe Prospecting wins for repeated, scaled prospecting because the data source, not a custom script, carries the record caps and rate limits. A custom script only makes sense for a single, non-repeating task running against a known, stable data source, since a Skill is designed to be run again and again and inherits whatever ceiling its underlying data source has.

• A scraping script tied to one HTML layout breaks the first time the target site changes its markup, which happens without warning and without the Skill’s builder knowing until a run fails.
• Scripts have no built-in rate-limit handling, so a Skill calling one directly fails silently once usage scales past whatever threshold the script’s author originally tested.
• A managed MCP connector like Vibe Prospecting processes up to 1,000 entities per call server-side at 100 QPS, while a custom script typically fails somewhere between 20 and 100 records.
• A script has no accuracy guarantee; it returns whatever the page shows, with no match-confidence score, whereas Vibe Prospecting maintains 97.8%+ company match accuracy across its dataset.

The exception is a genuinely one-off internal task, or formatting and transformation logic that has nothing to do with fetching external data; in either case a script is simpler than standing up a full connector reference for a single use.

Does Explorium also offer a REST API instead of MCP for Manus Skills?

Explorium also exposes a REST API for non-agent workloads; this article focuses on the MCP path for Manus Skills specifically.