hermes-gpt
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| HERMES_GPT_OWNER_ACK | No | Must equal 'I_UNDERSTAND_THIS_CAN_MUTATE_MY_MACHINE' for owner tools (default: unset). | |
| HERMES_GPT_ENABLE_WRITE | No | Set to '1' to enable write file and patch tools (default: unset, hidden). | |
| HERMES_GPT_OPERATOR_LEVEL | No | Operator level: read_only, cron, skills, skills_config, workspace, owner (default: read_only). | |
| HERMES_GPT_ENABLE_TERMINAL | No | Set to '1' to enable terminal command execution (default: unset, hidden). | |
| HERMES_GPT_OPERATOR_ENABLED | No | Set to '1' to enable operator mode (default: unset, false). | |
| HERMES_GPT_ENABLE_MEMORY_WRITE | No | Set to '1' to enable memory add, replace, remove actions (default: unset, disabled). | |
| HERMES_GPT_OPERATOR_APPLY_MODE | No | Apply mode: dry_run or direct (default: dry_run). | |
| HERMES_GPT_UNSAFE_REMOTE_NOAUTH | No | Set to '1' to bypass remote profile block for unsafe experiments (default: unset). | |
| HERMES_GPT_ENABLE_SESSION_SEARCH | No | Set to '1' to enable session search tool (default: unset, hidden). | |
| HERMES_GPT_OPERATOR_DENIED_PATHS | No | Extra denied paths (additions only, cannot weaken defaults; default: built-in). | |
| HERMES_GPT_OPERATOR_ALLOWED_PATHS | No | Comma-separated workspace root paths; empty disables workspace writes (default: empty). | |
| HERMES_GPT_OPERATOR_ALLOWED_PROFILES | No | Comma-separated profile names allowed, or '*' for all (default: default). |
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
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 43 tools
Each tool name follows a distinct 'resource_action' pattern, making their purposes clear despite most descriptions being empty. For example, 'hermes_cron_list' vs 'hermes_cron_run' are unambiguous.
All 43 tools use the consistent pattern 'hermes_<category>_<action>', with verbs in imperative form and underscores separating multi-word actions. No mixing of conventions.
With 43 tools, the server is overloaded. While each category has multiple tools, the overall count exceeds the typical well-scoped range, making it difficult for agents to navigate.
The tool surface covers many aspects (config, cron, env, skills, etc.) but has notable gaps like missing CRUD for cron (no create/delete) and environment (no create/delete), leaving agents without complete lifecycle coverage.