$ jcode provider-doctor conifer --model gpt-5.6-luna --tier offline --no-update
Provider doctor: Conifer / gpt-5.6-luna
Tier: offline (no API key, no spend: validates jcode wiring only)
Each line is one strict checkpoint. PASS/FAIL exercise it; skip means the
current tier does not run it (use --tier full for the API-dependent ones).

  [ skip] Credential loaded                      offline tier: no credential required
  [ skip] Live model catalog endpoint            offline tier: using synthetic catalog (no network)
  [ FAIL] Catalog hot reload in current session  

**Auth Model Catalog Warning**

Expected selectable Conifer model routes after auth, but found 2 matching route(s). Selected model: `gpt-5.6-luna`. Matching route sample: `fixture-model` via openai-compatible:conifer, `conifer-alternate-fixture-model` via openai-compatible:conifer.
  [ FAIL] Picker shows live models               selected model `gpt-5.6-luna` not present in picker entries
  [ PASS] Picker fallback labeling               all routes backed by live catalog (no static fallback)
  [ PASS] Model switch route                     switch request `conifer:fixture-model` routed via `openai-compatible:conifer`
  [ skip] Non-streaming chat completion          offline tier: requires --tier full (spends balance)
  [ skip] Streaming chat completion              offline tier: requires --tier full (spends balance)
  [ skip] Tool-call parse                        offline tier: requires --tier full (spends balance)
  [ skip] Tool execution loop                    offline tier: requires --tier full (spends balance)
  [ skip] Tool-result followup                   offline tier: requires --tier full (spends balance)
  [ skip] Real Jcode tool smoke                  offline tier: requires --tier full (spends balance)
  [ skip] Reasoning capability                   offline tier: requires --tier full (spends balance)

Verdict: FAILED at `Catalog hot reload in current session`.
  Reason: 

**Auth Model Catalog Warning**

Expected selectable Conifer model routes after auth, but found 2 matching route(s). Selected model: `gpt-5.6-luna`. Matching route sample: `fixture-model` via openai-compatible:conifer, `conifer-alternate-fixture-model` via openai-compatible:conifer.
  Next: this is a jcode-side routing/picker bug for this provider. Please file an issue with this output.

Error: provider-doctor: one or more checks failed for conifer
