Skillserver is dying after start

[fanningert]

I configurated skill server as standalone docker. I get only follwoing messages, when I enabled logging. After this message the docker instance is stopped with the status "exited".

2026-02-20T06:43:29.801654940Z 2026/02/20 06:43:29 Git syncer started
2026-02-20T06:43:29.802370700Z 2026/02/20 06:43:29 Starting web server on :8080

  localai-mcp-skillserver:
    container_name: localai-mcp-skillserver
    image: ghcr.io/mudler/skillserver:latest
    restart: unless-stopped
    ports:
      - "38083:8080"
    environment:
      SKILLSERVER_ENABLE_LOGGING: "true"
      SKILLSERVER_DIR: "/app/skills"
      SKILLSERVER_PORT: 8080
    volumes:
      - /mnt/user/appdata/skillserver/skills:/app/skills

[htarnacki]

I have the same problem ;-( It looks like the docker image doesn't work. There should definitely be some logging into the log file to be able to assess what is happening during the startup of the application/container

ok after some investigation i see the problem is the lack of pseudo-TTY. So basically adding -t resolves the problem:

    docker run -d -t \
      --name $CONTAINER_NAME \
      -p $PORT:$PORT \
      -e SKILLSERVER_DIR=/app/skills \
      -e SKILLSERVER_PORT=$PORT \
      -e SKILLSERVER_ENABLE_LOGGING='true' \
      -v "$SKILLS_PATH":/app/skills \
      --restart unless-stopped \
      $IMAGE

[CD-IE]

I have been getting the same issue, adding stdin_open: true and tty: true to my docker-compose file did not fix the issue. Also a stand alone docker container. Is there anything else I could try?