Keep tool-call assistant messages valid across provider sanitization and avoid trailing user-only history after model errors. This prevents follow-up requests from sending broken tool chains back to the gateway.
Some LLM providers (OpenAI-compat, Azure, vLLM, Ollama) reject requests
with consecutive same-role messages or trailing assistant messages. Add
_enforce_role_alternation() to merge consecutive same-role user/assistant
messages and strip trailing assistant messages before sending to the API.
Xubin Ren
Ziyan Lin