Skip to content

Identify whether Zed Agent or an external agent owns Montage

A working Montage entry in Zed settings does not automatically configure an external ACP agent or a terminal assistant.

Integrations

How do I identify whether Zed Agent or an external agent owns Montage?

A working Montage entry in Zed settings does not automatically configure an external ACP agent or a terminal assistant. Trace the caller before editing context_servers so the fix belongs to the surface that actually needs remote tools. Identify the actual assistant surface before choosing a configuration owner. Zed Agent context_servers, external ACP agents and terminal assistants can inhabit one editor while using independent registration and account authorization paths.

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 identify whether Zed Agent or an external agent owns Montage. Determine whether the affected conversation is Zed Agent, an external ACP agent or a terminal-based assistant. Record which agent returned signals_status and which surfaces remain unverified. No writes.
01 / Integrations

Identify the caller

Determine whether the affected conversation is Zed Agent, an external ACP agent or a terminal-based assistant. Sharing one editor window does not establish shared server registration. The ownership note should name the actual assistant surface, avoiding a broad “Zed failed” label that conflates three independently configured callers.

02 / Integrations

Inspect editor-owned settings

Review context_servers.montage.url only if Zed Agent owns the call. Preserve this definition if the failing caller is an independently configured external agent. A context_servers definition is evidence about editor-owned registration; it does not specify the external agent’s native MCP configuration file or account state.

03 / Integrations

Consult the actual owner

For an ACP or terminal caller, use that agent’s native MCP documentation instead of assuming Zed forwards its consent. Do not copy authentication storage between these surfaces. The external agent’s documented auth mechanism remains authoritative, and Zed’s browser consent should not be exported as a substitute for that client’s own access.

04 / Integrations

Test the intended conversation

Request a status read only from the identified owner after supported authorization. A successful Zed Agent result cannot validate an external agent’s distinct transport and account path. Use the caller’s actual execution result rather than a neighboring thread’s output so surface-specific proof remains attributable to the right session.

05 / Integrations

State surface-specific proof

Record which agent returned signals_status and which surfaces remain unverified. Avoid describing the entire editor as connected when only one independently configured caller has proven workspace access. The handoff should list the working caller explicitly and retain any ACP or terminal-path uncertainty instead of claiming the whole editor inherited access.

Keep the boundaries clear.

  • The external agent’s documented auth mechanism remains authoritative, and Zed’s browser consent should not be exported as a substitute for that client’s own access.
  • The handoff should list the working caller explicitly and retain any ACP or terminal-path uncertainty instead of claiming the whole editor inherited access.

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