Skip to content
Add doctor warnings for hidden MCP server tools | My AI Guide
App UpdateOpenClawv2026.5.20Vibe BuilderDeveloper

Add doctor warnings for hidden MCP server tools

By Harsh Desai
Share

TL;DR

The doctor utility now warns users when sandbox tool policies hide configured Model Context Protocol (MCP) server tools before provider requests are made.

What changed

The doctor utility in OpenClaw now warns users when sandbox tool policies hide configured Model Context Protocol server tools. The check runs before any provider requests are sent. This addresses cases where mcporter exposes MCP servers for databases or APIs but policies block them from view.

The update appeared in the May 2026 maintenance cycle for the self-hosted agent. It requires no new configuration flags and activates on the next doctor run after a CLI update.

Why it matters

Vibe Builders who connect external MCP tools gain an early signal instead of discovering blocks only after a failed task. The change reduces silent drops in automation flows that rely on third-party data sources.

It also pressures competing agents to surface similar hidden-tool diagnostics. OpenClaw bets that reliable MCP access will keep self-hosted users from migrating to managed platforms when they need custom tool servers.

How to use it

Update OpenClaw through the existing CLI command and run the doctor utility on your local install. Review the output for any MCP warnings tied to sandbox policies listed in your YAML file.

The feature works on current stable builds with no extra plan required. Test it after adding a new mcporter bridge to confirm the warning triggers correctly.

Watch for

Watch for reduced runtime errors on MCP calls as confirmation the check works. It fails if policy syntax changes faster than the doctor parser. The next logical step is similar pre-checks for browser and shell tool visibility.

Who this matters for

  • Vibe Builders: Run the doctor utility after adding mcporter bridges to catch hidden tool blocks before tasks fail.
  • Developers: Use the updated doctor CLI to debug sandbox policy conflicts that silently drop MCP server access.

Harshs take

Silent failures in MCP tool discovery are a major friction point for self-hosted agent stacks. OpenClaw is making a smart move by moving these checks to the doctor utility. This shift from runtime error to pre-flight diagnostic saves significant debugging time when bridging external databases or APIs via mcporter.

This update signals a focus on reliability for complex, multi-tool environments. By surfacing sandbox policy blocks before provider requests occur, OpenClaw reduces the trial and error cycle for local setups. It is a pragmatic improvement that prioritizes system transparency over flashy features, ensuring that custom tool servers actually work when called.

by Harsh Desai

Source:myaiguide.co

About OpenClaw

View the full OpenClaw page →All OpenClaw updates

More from OpenClaw

Everything AI. One email.
Every Monday.

New tools. Model launches. Plugins. Repos. Tactics. The moves the sharpest builders are making right now, before everyone else.

No spam. Unsubscribe anytime.