Skip to content
TheChatArchiveThe Chat Archive
composer-2.5· unknownHugging Face ibragim-bad/swe_rebench_07_2026_trajectories

ssh Tunnel Model Transfer issue I have installed hermes in a Docker Container on Ubuntu…

  1. User

    ssh Tunnel Model Transfer issue I have installed hermes in a Docker Container on Ubuntu Server and Connection it with a Customer Router to a free nividea Neutron 120B Modell and it worked over Telegramm. (I used squid for blocking Special Download requests or webpages) Then I installed the Desktop App on Windows my Main pc, started it and connected via ssh but inside i just couldnt See the Model I Like „Coded in“ so I tried to Connect to via App but everytime I Go through the process of tipping everything it Lords and then Colosseum but the Modell doesnˋt Sems to change or Save.

    Evaluation contract

    Remote YAML lookup must resolve dotted paths only through direct children inside the selected indentation block and must not cross into sibling blocks or match a same-named key under another parent. Flat-key lookup is restricted to column zero. Return null for missing parents or leaves, strip matching single or double quotes from returned values, and report offsets that point exactly to the raw value substring so writers can replace it safely. Support arbitrary nesting depth.

    Interface

    Type: Interface Name: YamlPathHit Path: src/main/ssh-remote.ts Fields: value: string, valueStart: number, valueEnd: number Description: Result returned by YAML key lookup, containing the unquoted value and its exact source offsets.

    Type: Function Name: findTopLevelKey Path: src/main/ssh-remote.ts Input: content: string, key: string Output: YamlPathHit | null Description: Resolves a key at column zero without matching nested occurrences.

    Type: Function Name: findYamlPath Path: src/main/ssh-remote.ts Input: content: string, dottedPath: string Output: YamlPathHit | null Description: Resolves a dotted YAML path while respecting parent-child indentation boundaries.

  2. Assistant

    Investigating SSH remote model transfer and YAML path lookup — likely why the configured model isn't visible or saving on the Desktop app.

  3. Assistant

    Implementing YAML path lookup and updating SSH config read/write so nested keys like models.default resolve correctly instead of matching the wrong provider elsewhere in the file.

  4. Assistant

    Adding tests for the YAML path lookup functions.

  5. Assistant

    I tracked this down to a bug in how the Desktop app reads and writes config.yaml over SSH.

    Root cause

    Hermes stores model settings under a nested models: block:

    models:
      default: "nvidia/neutron-120b"
      provider: openrouter
      base_url: http://your-router/v1
    

    The SSH remote code used a flat regex that matched the first provider: or default: anywhere in the file. So if you also had something like:

    memory:
      provider: honcho
    models:
      default: "nvidia/neutron-120b"
      provider: openrouter
    

    …the app could read honcho as the model provider, show the wrong model, and fail to save changes to the correct keys.

    Fix

    I implemented proper YAML path lookup in ssh-remote.ts:

    • findTopLevelKey — only matches keys at column 0 (e.g. legacy flat provider:)
    • findYamlPath — resolves dotted paths like models.default and memory.provider within the correct parent block
    • sshGetModelConfig / sshSetModelConfig — prefer models.* keys, with fallback to top-level keys for older configs

    Model reads and writes over SSH should now target the right values in your Docker config.yaml.

    What to do on your side

    1. Reconnect SSH in Settings (or restart the Desktop app) so it picks up the fix.
    2. Confirm remote config on the server:
      docker exec -it hermes cat /opt/data/config.yaml
      
      (Path may differ — check where ~/.hermes is mounted in your container.)
    3. Set the model again in the Desktop app Providers/Models screen — it should now persist to models.default, models.provider, and models.base_url.
    4. Restart the gateway if Telegram still uses the old model:
      docker exec -it hermes hermes gateway restart
      

    Note on the models list

    Over SSH, the saved-models list comes from remote ~/.hermes/models.json. If that file is empty, the list can look bare even when config.yaml has the right active model. The active model in chat should still show correctly after this fix. If you want your custom Neutron model in the saved list, add it once in the Models screen (after reconnecting) or add it to models.json on the server.

    If the app still closes after connecting, check ~/.hermes/logs/agent.log on the server or the Desktop app logs — that may be a separate SSH/auth issue. Share those logs if it keeps happening.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.