Skip to content

Constrain VS Code Montage tool selection before workspace changes

VS Code can recognize an MCP server without enabling its tools in the conversation that needs them.

Integrations

How do I constrain VS Code Montage tool selection before workspace changes?

VS Code can recognize an MCP server without enabling its tools in the conversation that needs them. Set a read-only checkpoint in the affected chat, so that enabling a server does not silently broaden a coding task into remote research mutations. The status request belongs in the affected chat after its tool selection is reviewed. Selecting a server does not itself approve watchlist creation, audience mutations or automatic continuation of an earlier coding instruction.

01 / CONFIGUREOne remote endpointUse this client's documented HTTP setup
02 / AUTHORIZEYour workspace accessUse a supported sign-in or credential flow
03 / VERIFYRun signals_statusCheck a real workspace response
Copyable task-specific request
Help me constrain VS Code Montage tool selection before workspace changes. Start with the chat where Montage should be available, not another window that already works. Summarize the verified read or the remaining tool-selection blocker and stop. No writes.
01 / Integrations

Identify the affected conversation

Start with the chat where Montage should be available, not another window that already works. Explicitly distinguish local coding permissions from permission to change remote watchlists or audiences. A conversation-specific tool inventory helps distinguish an unavailable tool from an available call that later fails because authorization was rejected.

02 / Integrations

Review tool selection controls

Use the installed VS Code assistant’s documented tool controls to enable the intended server. A server entry alone does not prove that this conversation can select its status tool. Use documented assistant controls for the installed version rather than assuming another extension’s tool picker or permission behavior applies here.

03 / Integrations

Name the only allowed execution

Ask for signals_status and keep all write proposals unexecuted. If the assistant offers another operation as a connectivity substitute, reject that expansion and retain the narrower diagnostic. A narrow request should also prohibit substitute mutation tests, since creating records is not required to show that account-scoped status can execute.

04 / Integrations

Inspect the returned workspace

Compare real status output with the expected account and available coverage. A tool-picker checkbox establishes selection, whereas only the response establishes this conversation’s authorized workspace read. Where response coverage is limited, retain the limitation without interpreting it as zero data or an instruction to expand the diagnostic.

05 / Integrations

End at the approval boundary

Summarize the verified read or the remaining tool-selection blocker and stop. Any future workspace mutation needs an explicit decision rather than inheriting approval from a successful editor-side status check. The handoff can name the next proposed write while leaving it pending; describing an action does not require executing it to explain scope.

Keep the boundaries clear.

  • A narrow request should also prohibit substitute mutation tests, since creating records is not required to show that account-scoped status can execute.
  • The handoff can name the next proposed write while leaving it pending; describing an action does not require executing it to explain scope.

Start with a read-only check: ask your agent to report actual source coverage before collecting records or exporting an audience.

Primary references

Documentation reviewed 2026-11-15. Product capabilities, client availability and plans can change. Third-party marks identify their products, not an affiliation.

Know what social media shows about you.
Start with a question.

Get started
montage