DataMerge

Enrich companies and find B2B contacts using the DataMerge MCP server (mcp.datamerge.ai). Use when the user needs company firmographic data, validated contac...

Security Scan

The skill's behavior matches its description (DataMerge API usage) and is instruction-only, but the registry metadata omits declaring the required DataMerge API credential which reduces confidence slightly.

Purpose & Capability

Name/description align with the instructions: all steps reference the DataMerge MCP server (mcp.datamerge.ai) and DataMerge workflows (company enrichment, contact search, lookalikes, lists). There are no unrelated services, binaries, or config paths requested.

Instruction Scope

SKILL.md contains clear, scoped runtime instructions for calling the DataMerge API, polling jobs, and using credits. It does not instruct reading local files, unrelated env vars, or transmitting data to unexpected endpoints beyond app.datamerge.ai and mcp.datamerge.ai.

Install Mechanism

No install spec and no code files — the skill is instruction-only, so nothing is downloaded or written to disk by an installer.

Credentials

The instructions require a DataMerge API key (explicitly mentioned), but the registry metadata lists no required env vars or primary credential. This is a minor inconsistency: the skill will need user-provided credentials at runtime even though none are declared in metadata.

Persistence & Privilege

always:false and no config paths requested. The skill does not request persistent system presence or elevated privileges. Model invocation is allowed (default), which is expected for a service integration.

Assessment

This skill appears to do what it says: it calls DataMerge APIs and requires the user's DataMerge API key. Before installing or using it: (1) Confirm you trust https://mcp.datamerge.ai and https://app.datamerge.ai and review their privacy/terms for sending contact/company data; (2) Note the registry metadata did not declare the API key — be clear how the key will be provided/stored and avoid entering a high-privilege or production key until tested; (3) Prefer using a limited/test API key or account and monitor credit usage; (4) Be cautious about sending sensitive PII (personal emails/phone numbers) to the service; (5) If you do not want the agent to call the external API autonomously, disable autonomous invocation or require user consent before actions. If you want higher assurance, ask the publisher to add a declared primary credential in the metadata and provide a privacy/data-handling statement.

Connect to the DataMerge MCP server at https://mcp.datamerge.ai.

Auth

Call configure_datamerge with the user's API key before using any other tool. New users get 20 free credits at https://app.datamerge.ai.

Credits

• 1 credit: company enrichment, validated email
• 4 credits: mobile phone number
• record_id retrieval (get_company, get_contact): free — always use this to re-fetch

Core workflows

Enrich a company

Use start_company_enrichment_and_wait for single domains — it polls automatically and returns when complete. Use start_company_enrichment + get_company_enrichment_result for batch jobs.

Find contacts

1. contact_search with target domains and enrich_fields: ["contact.emails"]
2. Poll get_contact_search_status until completed
3. get_contact with each record_id to retrieve details (free)

Use job_titles to filter by seniority. Start with emails only — add "contact.phones" only if mobile numbers are explicitly needed (4× the cost).

Find lookalike companies

1. start_lookalike with companiesFilters.lookalikeDomains (seed domains)
2. Poll get_lookalike_status until completed
3. get_company with each record_id (free)

Company hierarchy

Enrich the company first to get a datamerge_id, then call get_company_hierarchy. Set include_names: true to get entity names (costs 1 credit).

Lists

• create_list to save a group of companies or contacts
• Pass list slug to enrichment/search jobs to add results automatically
• get_list_items to retrieve saved records
• Use skip_if_exists: true to avoid re-enriching duplicates

Tips

• Check get_credits_balance before running large batch jobs
• global_ultimate: true returns the top-level parent company instead of the subsidiary
• strict_match: true requires an exact domain match — use when precision matters over coverage