hermes-agent-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| HERMES_HOME | No | Hermes data directory | ~/.hermes |
| HERMES_MCP_BIN | No | Path to the CLI | hermes on PATH |
| HERMES_MCP_TIMEOUT | No | Seconds for ordinary CLI calls | 120 |
| HERMES_MCP_MAX_OUTPUT | No | Characters returned per call before truncation | 40000 |
| HERMES_MCP_ALLOW_SHELL | No | Enable the shell tool | 0 |
| HERMES_MCP_ALLOW_WRITE | No | Allow tools that change state (cron edit, config set, restart). Set 0 for read-only | 1 |
| HERMES_MCP_ASK_TIMEOUT | No | Seconds for hermes_ask | 900 |
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
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| hermes_statusA | Report how this server is configured and whether the Hermes CLI answers. |
| hermes_askA | Give the Hermes agent a task and return its reply. Runs one non-interactive turn with the agent's own tools, skills and memory. Use this for work; use the specific tools below for inspection and config. |
| doctorA | Run |
| gateway_statusA | Show whether the messaging gateway is running and supervised. |
| gateway_restartA | Restart the messaging gateway so configuration changes take effect. |
| cron_listA | List scheduled jobs with their schedules and next run times. |
| cron_createB | Create a scheduled job.
|
| cron_editB | Change a scheduled job in place, keeping its id and run history. |
| cron_runA | Queue a job to run on the next scheduler tick instead of waiting. |
| cron_runsA | Show recorded execution attempts, newest first. |
| config_getA | Read one config.yaml value by dotted key, e.g. |
| config_setA | Write one config.yaml value with an explicit type.
|
| skills_listA | List installed skills with category, source and enabled state. |
| sessions_listA | List recent sessions so you can resume or inspect one. |
| model_infoA | Show the configured default model and provider, plus Nous Portal status. |
| shellA | Run a shell command on the Hermes host. Disabled unless HERMES_MCP_ALLOW_SHELL=1. Intended for agents that can reach this MCP server but have no terminal of their own. |
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 16 tools
Most tools target distinct concerns (cron vs config vs gateway), but hermes_status, doctor, config_get, and model_info all provide overlapping status/configuration information, requiring careful description reading.
Naming patterns are mixed: single nouns (doctor, shell), noun_noun compounds (hermes_status, model_info), and noun_verb constructions (cron_create, config_get) appear without a consistent convention. Some tools have domain prefixes while others do not.
At 16 tools, the set sits at the heavy end of the reasonable range. The breadth of domains (agent execution, gateway, cron, config, skills, sessions, shell) mostly justifies the count, but it feels slightly sprawling for a single server.
Core agent interaction and inspection are well covered, and cron has rich lifecycle support, but notable gaps exist: no cron_delete, no skill management beyond listing, and session support stops at listing with no resume or detail operations.