extrovert
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| EXTROVERT_HOME | No | Overrides the state directory used for extrovert's data files (config.json, trust.json, broker.log, logs/). Defaults to ~/.extrovert. | |
| EXTROVERT_MOUNT_TIMEOUT | No | Timeout in seconds for the first mount (default 90). Useful because the first download of an npm package can be slow. | 90 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| capability_searchA | Search the official MCP Registry for installable MCP servers. Returns structured candidates with name, description, version and packages (npm/pypi identifiers). Pass a server's name to capability_add to install it. |
| capability_considerA | Planning-phase advice: given a subtask description, suggest existing capabilities from the official MCP Registry that could do it (with local verification status). Call while decomposing a task, before implementing anything yourself. Results are cached for 10 minutes; registry outages degrade to 'implement it yourself' advice. Use capability_search for raw candidate lists. |
| capability_addA | Install and mount an MCP server. identifier is either a registry server name (e.g. 'com.example/foo', see capability_search) or a local stdio command (e.g. 'python3 /path/to/server.py'). The server's tools become available immediately as '__'. |
| capability_verifyA | Smoke-test a mounted server: ping, list_tools, and one safe zero-argument probe call (echo/ping/hello/version) when available. The result is recorded in the local trust store. This is a liveness check, not a security audit. |
| capability_listA | List mounted capability servers with their tools, source, health and verification status. |
| capability_removeA | Unmount a capability server: terminate its subprocess, unregister its '__' tools and remove it from the persisted config. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 6 tools
Each tool maps to a distinct lifecycle step (search, consider, add, verify, list, remove), but capability_search and capability_consider both query the official MCP Registry and could be confused; the descriptions do clarify that consider gives planning advice with verification status while search returns raw candidates.
All six tools share the same capability_<verb> snake_case pattern with clear, single-word verbs (search, consider, add, verify, list, remove).
Six tools is a well-scoped set that exactly covers the discovery-to-removal lifecycle of mounting MCP servers, with no redundant entries.
The mount lifecycle is fully covered (find, plan, install, smoke-test, inspect, unmount), but there is no operation to update/reconfigure an already-mounted server or its alias, and the trust store written by capability_verify is only visible indirectly via capability_list.