ur-agent 1.27.5 → 1.28.0

package/CHANGELOG.md

@@ -1,5 +1,35 @@
 # Changelog
 
+## 1.28.0
+
+- Subscription CLI providers (Codex, Claude Code, Gemini, Antigravity) now
+  perform real dispatch: the official CLI is spawned in non-interactive mode with
+  the scoped model and prompt, and its stdout becomes the response. Non-zero exit
+  or empty output fails clearly instead of returning placeholder text.
+- API providers now use each provider's native wire format: Anthropic
+  `x-api-key` + `anthropic-version` on `/v1/messages`, OpenAI `Authorization:
+  Bearer` on `/v1/chat/completions`, Gemini `x-goog-api-key` on
+  `:generateContent`, routed by a new provider-family classifier.
+- Real runtime provider identity: `getRuntimeProviderId`/`getProviderFamily`
+  expose the true selected provider; `getAPIProvider` is derived from it rather
+  than string-matching raw settings.
+- Fix saved local/server (live-discovery) model pairs being rejected on a cold
+  process; the endpoint is treated as the source of truth before discovery runs.
+  Static (API/subscription) pairs remain strictly validated.
+- Add behavior-proving tests that exercise the real clients (CLI runner, mocked
+  HTTP) and assert wire format, response content, identity, and cold-cache
+  restart — not just provider-id routing. Centralize the Ollama default and
+  remove dead code.
+
+## 1.27.6
+
+- Route runtime requests through the selected provider/model pair instead of
+  allowing stale Ollama/default-provider paths to handle non-Ollama requests.
+- Add runtime dispatch validation, backend labels, and focused mocked dispatch
+  tests for subscription CLI, API, Ollama, and OpenAI-compatible providers.
+- Fix malformed message rendering from crashing startup when an undefined entry
+  appears in the UI message list.
+
 ## 1.27.5
 
 - Make `/model` provider-first and provider-scoped, with clear subscription,

package/README.md

@@ -143,9 +143,11 @@ ur --ollama-host http://192.168.1.50:11434
 ur --discover-ollama
 ```
 
-Model selection precedence is `OLLAMA_MODEL`, then `UR_MODEL`, then the model
-router over the configured Ollama host. If model discovery fails, the built-in
-fallback is `qwen3-coder:480b-cloud`.
+For Ollama sessions, model selection precedence is `OLLAMA_MODEL`, then
+`UR_MODEL`, then the model router over the configured Ollama host. Once a
+provider/model pair is selected with `/model` or `ur config set`, runtime
+requests use that provider backend; they do not fall back to Ollama unless
+`ollama` is the selected provider.
 
 ### Legal Provider Auth
 
@@ -195,7 +197,10 @@ Use quotes for shell values with spaces.
 In the interactive app, `/model` shows providers first and then shows only the
 selected provider's models. OpenAI API and Codex CLI, Claude API and Claude
 Code, and Gemini API and Gemini CLI are separate access paths; subscription
-logins and API keys are not interchangeable.
+logins and API keys are not interchangeable. The final confirmation includes
+the runtime backend that will receive the next agent request. Use
+`ur provider status` to show the active provider, model, access type, and
+runtime backend.
 
 Security policy: UR-AGENT never scrapes browser sessions, extracts OAuth
 tokens, bypasses subscription/quota/region/organization restrictions, proxies a