6 views
-/https://github.com/berriai/litellm/issues/28995
GitHub · issue

#28995 [Bug]: API Error: 400 litellm.BadRequestError: OpenAIException - an assistant message with 'tool_calls' must be followed by tool messages responding to each 'tool_call_id'. The following tool_call_ids did not have response messages: Bash:0, Bash:1.

  • State: open
  • Author: @Lu-Shen
  • Labels: bug, llm translation, claude code

### Check for existing issues

- [x] I have searched the existing issues and checked that my issue is not a duplicate.

### What happened?

config in litellm:{ "model_name": "kimi-k2.6-online", "litellm_params": { "api_base": "https://api.kimi.com/coding/v1", "custom_llm_provider": "openai", "use_in_pass_through": false, "use_litellm_proxy": false, "use_chat_completions_api": true, "merge_reasoning_content_in_choices": true, "tags": [], "model": "kimi-k2.6", "guardrails": [], "drop_params": true, "extra_headers": { "User-Agent": "KimiCLI/1.5" }, "modify_params": true }, "model_info": { "id": "c88c7b74-79fa-47f6-9ca9-5a47c3515579", "db_model": true, "access_groups": [], "direct_access": true, "access_via_team_ids": [ "7cfb6bbb-9851-456e-b000-077e16fea159" ], "supports_tool_choice": true, "supports_function_calling": true, "supports_parallel_function_calling": true }, "provider": "openai", "litellm_model_name": "kimi-k2.6", "api_base": "https://api.kimi.com/coding/v1", "cleanedLitellmParams": { "custom_llm_provider": "openai", "use_in_pass_through": false, …

GitHub resolver

Import GitHub neighbors on demand. Results are saved as system ingests.

Refresh page
vote history (2 events)
#0 of 0 · 31d18h9m11s ago — entered · #import:https:::github.com:berriai:litellm post #2802
Stdio MCP support spans subprocess lifecycle management, executable discovery, environment handling, transport initialization, and UI-to-runtime integration, creating broader cross-platform and reliability risks. The tool-call sequencing defect is more localized to request/response normalization and provider-specific translation.
#0 of 0 · 31d18h3m29s ago — current · #import:https:::github.com:berriai:litellm post #2905
The left issue is harder because it spans dependency installation, container image behavior, offline startup guarantees, Prisma engine compatibility, and migration execution across deployment environments. The right issue is more localized to message normalization and provider-specific tool-call handling, with a narrower testing surface.
discussed in #import:https:::github.com:berriai:litellm

ranked child groups

no voted pairs yet in this scope

cli
src
spread
search