hermes-gpt
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| HERMES_GPT_SAFE_ROOT | No | The only directory where write tools may create files. | |
| HERMES_GPT_HTTP_TOKEN | No | Enables query/Bearer token enforcement on HTTP/SSE transports. | |
| HERMES_GPT_READ_ROOTS | No | os.pathsep-separated read-only roots visible to read/search/tree/image tools. | |
| HERMES_GPT_ENABLE_WRITE | No | Set to '1' to enable create-only sandbox write (hidden by default). | |
| HERMES_GPT_PATH_ALIASES | No | Public-to-local path aliases, e.g. '/linus1=/home/imac-hermes/linus1/shared'. | |
| HERMES_GPT_HOST_IDENTITY | No | Human label such as 'Dell' or 'Pedro' returned by hermes_server_config. | |
| HERMES_GPT_ENABLE_TERMINAL | No | Set to '1' to enable terminal command execution (hidden by default). | |
| HERMES_GPT_REQUIRE_HTTP_TOKEN | No | Set to '1' to reject HTTP traffic if no token is configured. | |
| HERMES_GPT_ENABLE_MEMORY_WRITE | No | Set to '1' to enable memory add/replace/remove (disabled by default). | |
| HERMES_GPT_UNSAFE_REMOTE_NOAUTH | No | Set to '1' to bypass remote-profile guard for no-auth experiments (unsafe). | |
| HERMES_GPT_ENABLE_SESSION_SEARCH | No | Set to '1' to enable session search (hidden by 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
| Name | Description |
|---|---|
| hermes_server_configA | Show Hermes GPT bridge identity, allowed read roots, write sandbox, auth status, and safety modes. Does not reveal tokens. |
| hermes_self_testA | Run a local bridge self-test for host identity, auth, read roots, write sandbox, and sensitive path blocking. |
| hermes_treeA | Show a shallow directory tree for an allowed local path. Read-only and safer than regex search for discovery. |
| hermes_read_fileB | Read a local file from this Hermes machine. Use only when the user provides an explicit path. |
| hermes_search_filesB | Search allowed local files on this Hermes machine. For filenames use target='files' or target='path'. Read-only. |
| hermes_list_filesA | List files under an allowed local directory. Read-only; safer than search when the user asks what files are visible. |
| hermes_view_imageA | View an allowed local image file as image content. Use this for PNG/JPEG/WebP/GIF design review. |
| hermes_skill_listA | List installed Hermes skills on this machine. Read-only and safe for a connector smoke test. |
| hermes_skill_viewA | Read the SKILL.md content for one installed Hermes skill by name. Read-only. |
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 9 tools
Each tool has a clearly distinct purpose: file listing, reading, searching, self-test, config, skills, directory tree, and image viewing. No overlap or ambiguity.
All tools follow a consistent 'hermes_verb_noun' pattern in snake_case, making them predictable and easy to understand.
9 tools is well-scoped for a server focused on safe local file access, skill management, and configuration. Each tool serves a necessary function without redundancy.
Coverage is strong for read-only operations and skill management. The only gap is the lack of write/delete capabilities, but this is intentional for safety, making the set complete for its design purpose.