Badger Maps (Standard) MCP server icon

Badger Maps (Standard)

by Badger Maps (Standard)

HIPAA CompliantSOC2 ReadyISO 27001 Ready
Productivity21 tools

Read and change a field-sales territory from an AI agent — accounts, routes, check-ins and follow-ups. Anthropic lists 21 tools: 13 read, 7 write, 1 delete. One opaque OAuth scope named badger, and per-tool permissions in the client are the only real boundary.

Verified connector

Listed by Anthropic as a partner connector in its Connectors Directory.

Connection checked by Agentman on .

Anthropic states this reflects the level of review a connector received, not a security audit.

Connect Badger Maps (Standard) via MCP

https://mcp.badgermapping.com/mcp

Works in any MCP-compatible client. In Agentman Studio it is one click — no config file to edit.

Badger Maps (Standard) Tools & Capabilities (21)

Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.

Limits

  • The single OAuth scope draws no boundary. badger names no object and no verb. Nothing at the consent screen separates the 13 reads from the seven writes or the one delete. The only place that separation exists is the client's per-tool permission setting.
  • Eleven of the 21 tool names have no vendor corroboration in any medium — no text, no recording. That set includes delete_account and six of the seven writes. The ten names Badger's recordings do corroborate are nine reads plus create_route, and they match as prettified labels rather than as literal strings, so a rename would not have shown.
  • delete_account is undocumented. Badger publishes no description, no statement on whether the deletion is recoverable, and no mention of a confirmation gate. We did not test it.
  • We could not read tool schemas or safety annotations. The endpoint returns 401 to an anonymous request, so no parameter-level detail appears on this page and every read/write label is our classification of a name.
  • The connector inherits the connecting user's clearances, and Badger names no MCP-specific limit. An admin login gives the agent admin breadth across the team's accounts, routes and check-ins.
  • The AI leg is your responsibility, and Badger says so. Its documentation states the user assumes full responsibility for compliance when connecting the server to any external LLM or AI platform, and that it makes no representations about third-party platforms' data practices. Where reps are tracked, employee-notification and works-council duties are the operator's.
  • Data may reach model training via the AI platform, not via Badger. Badger's documentation warns that data returned from the MCP server may be processed by the connected platform and, depending on account settings, used to improve or train its models, and directs users to that platform's opt-out. Badger notes this is governed by the third party's policy, not its own.
  • A second Badger MCP server exists at a different endpoint and is not in the current directory. See below — it is not what this page describes.
  • No metering signal either way. No tool name carries a credit, balance, usage, quota or billing term, and Badger publishes no per-call cost for MCP. We found no evidence that a call spends anything, and no tool that could observe it if one did.
  • Badger publishes no rate limits for the MCP server. Its security page describes login rate-limiting and account lockouts, which is a different control.

Frequently asked questions

It reaches their visit history, not their live position. get_checkins returns the check-in records a rep filed at customer sites, each carrying a place and a time, and list_routes plus get_route return the planned stops for a day. Ordered by time that is a movement log for a named worker. Badger's privacy policy says device location is captured non-continuously.

Yes, and they are the connector's main subject. Accounts in Badger are the businesses a rep visits, and list_accounts, search_accounts, get_account and get_accounts_in_area return them. Badger's demo transcripts show account names, street addresses, recipient email addresses, last check-in dates and free-text notes coming back. Those are third-party contact records, not your own data.

One tool can: delete_account. It is the only destructive name among the 21, and it removes a customer account record along with whatever history hangs off it. Badger publishes no documentation for the tool and we could not read its schema, so whether deletion is soft or permanent, and whether it cascades to check-ins, is unknown.

Exactly one, named badger. The server's RFC 9728 descriptor and its authorization server metadata both advertised that single scope on 2026-08-23. It names no object and no verb, so it cannot separate reading an account from deleting one. Granting it grants whatever the server implements, now and in future.

Use per-tool permissions in the client, because OAuth will not do it. Badger's guides state that in Claude you go to Customize, then Connectors, then Badger Maps, then Tool permissions, and set each tool to Always allow, Needs approval, or Blocked. That per-tool control is the only place the write and delete tools can be separated from the read ones.

Not by itself — no tool among the 21 sends mail. Badger's own use cases pair the connector with a separate Gmail or Outlook connector, and its example prompts all end with a saved draft rather than a send. In Badger's demo recordings Claude states the drafts are saved and the Send button is never clicked, so a human still presses send.

Exactly the ones the connecting user already has. Badger states the MCP server is a wrapper around its API and carries the same permissions as the setup user, and gives the example that connecting through an admin who can already see everything gives that same admin access over MCP. Connect a rep-level account, not an admin one, to narrow the reach.

Badger says the MCP server itself does not, but its wider product does log AI usage. Its documentation states the MCP server does not independently store, process, or share user or customer data with third parties. Separately its privacy policy names PostHog as storing AI Assistant chat transcripts and related metadata, which describes Badger's in-app assistant rather than this connector.

Sources

Use in Agentman

Connect once and your agents call these tools on their own — on a schedule, in a workflow, with nobody at the keyboard.

Open in Agentman Studio

Server Info

Category
Productivity
Developer
Badger Maps (Standard)
Tools
21
Domain
mcp.badgermapping.com

Using Claude Desktop or another MCP client? Setup docs — the connection URL above works anywhere.