herdr-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| HERDR_MCP_MACHINE | No | Target a remote machine saved in herdr (a saved profile label or ID, not a hostname). Every command is sent as `herdr --machine "<value>" ...`, which herdr forwards over SSH. Workspace, pane, and agent IDs belong to that machine. | |
| HERDR_MCP_ALLOW_WRITE | No | Set to exactly `1` to enable write tools. By default the server is read-only and the 6 write tools are not registered. Example: `"env": { "HERDR_MCP_ALLOW_WRITE": "1" }`. |
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": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| herdr_statusA | Check that the herdr server is running and report its version. Call this first if other herdr tools fail. |
| herdr_workspace_listA | List herdr workspaces. A workspace is a project-level container that owns tabs and panes and rolls up agent states. |
| herdr_tab_listA | List tabs. Tabs are layouts inside a workspace; each contains one or more panes. |
| herdr_pane_listB | List panes. A pane is a real terminal that survives client detach; it may hold a shell, a process, or a coding agent. |
| herdr_agent_listA | List the coding agents herdr recognizes, with each agent's status (idle/working/blocked/done/unknown) and pane. The primary monitoring tool. |
| herdr_agent_getA | Get one agent's status, pane location, working directory, terminal title, and session info. |
| herdr_agent_readA | Read recent terminal output from an agent's pane as plain text. Use it to see what an agent did, what it is asking, or why it failed. |
| herdr_agent_waitA | Wait until an agent reaches a status. Without 'until', waits for idle, done, or blocked. Returns immediately if the agent already matches. |
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 8 tools
Each tool has a distinct purpose: status check, lists for workspaces/tabs/panes, and agent-specific operations (list, get, read, wait). The list/read/wait agent tools are clearly differentiated by their descriptions, so an agent can easily choose the right one.
Almost all tools follow a consistent herdr_ prefix and noun_verb pattern (e.g., herdr_workspace_list, herdr_agent_get). The only deviation is herdr_status, which lacks a verb, but it remains readable and predictable overall.
8 tools is a well-scoped set for a monitoring-focused server. Each tool earns its place by covering a distinct resource or agent operation without redundancy.
The surface covers listing workspaces, tabs, panes, and agents, plus detailed agent monitoring (get, read, wait) and server status. However, it lacks get operations for workspaces, tabs, or panes, and offers no control actions (e.g., create, close, start agent), which are minor gaps for a monitoring server.