llm 0.31.1

The software tool LLM, which provides command-line access to large language models, has been updated to version 0.31.1. This release serves as a maintenance update to improve the reliability of the software when interacting with specific AI service providers.

The primary change in this version is a fix for a bug affecting OpenAI Chat Completion endpoints. Previously, when a tool call was made with empty arguments, some providers would return a JSON error. This issue was identified during testing of the llm-meta-ai plugin and has now been resolved to ensure smoother communication between the command-line interface and external models.

This update is important for developers who rely on the LLM tool for automated tasks and agentic workflows. By correcting how empty tool arguments are handled, the software maintains better compatibility with various API providers, preventing unnecessary errors during complex model interactions.