Shelley client create chat and default model

[nzoschke]

The example in https://github.com/boldsoftware/shelley/blob/main/client/client.go doesn't work for me.

# Start a conversation and wait for the agent
ID=$(shelley client chat -p "list files" | jq -r .conversation_id)
shelley client read -wait "$ID"
<omitted>

# Continue a conversation
shelley client chat -c "$ID" -p "now count them"
Error: HTTP 400

Digging in, it looks like the initial chat is created with gpt-oss-20b-fireworks, but continuing uses the default model. Specifying the right model to continue the chat works.

shelley client list
{"conversation_id":"cFA76EP","slug":"list-files-3","created_at":"2026-03-05T15:11:49Z","updated_at":"2026-03-05T15:12:01Z","working":false,"model":"gpt-oss-20b-fireworks"}

shelley client chat -c "$ID" -model "gpt-oss-20b-fireworks" -p "now count them"
{"conversation_id":"cFA76EP"}

Config on the VM

shelley version 
{
  "version": "0.326.913373364",
  "tag": "v0.326.913373364",
  "commit": "2df6f4a9f8764de87c013ba4ed1902a688879743",
  "commit_time": "2026-03-05T05:25:56Z",
  "modified": true
}

$ jq . /exe.dev/shelley.json  
{
  "default_model": "claude-opus-4.6",
  "key_generator": "echo irrelevant",
  "links": [
    {
      "icon_svg": "M3 12l2-2m0 0l7-7 7 7M5 10v10a1 1 0 001 1h3m10-11l2 2m-2-2v10a1 1 0 01-1 1h-3m-6 0a1 1 0 001-1v-4a1 1 0 011-1h2a1 1 0 011 1v4a1 1 0 001 1m-6 0h6",
      "title": "Back to exe.dev",
      "url": "https://exe.dev"
    },
    {
      "icon_svg": "M13.19 8.688a4.5 4.5 0 011.242 7.244l-4.5 4.5a4.5 4.5 0 01-6.364-6.364l1.757-1.757m13.35-.622l1.757-1.757a4.5 4.5 0 00-6.364-6.364l-4.5 4.5a4.5 4.5 0 001.242 7.244",
      "title": "nimble-tiger.exe.xyz",
      "url": "https://nimble-tiger.exe.xyz"
    }
  ],
  "llm_gateway": "http://169.254.169.254/gateway/llm",
  "terminal_url": "https://nimble-tiger.xterm.exe.xyz?d=WORKING_DIR"
}