multi-provider-llm-proxy-debugging

What problem does it solve?

This Skill helps diagnose and fix issues in multi-provider LLM API proxy chains where fallbacks fail, providers return silent errors, or error bodies are misinterpreted as responses.

Core Features & Use Cases

• Diagnose Silent Failures: Identify why a proxy chain isn't falling back correctly or why providers appear to succeed but return no usable data.
• Resolve Error Body Misinterpretation: Ensure that actual error messages from LLM providers are not parsed as valid responses.
• Fix URL Path Construction: Correct issues where base URLs are incorrectly concatenated, leading to 404 errors.
• Handle Provider-Specific Fields: Adapt requests to comply with the varying strictness of different LLM providers' API layers.
• Improve Error Visibility: Ensure error bodies are readable in logs by managing compression headers.
• Use Case: When your bot starts returning empty responses or errors from an LLM proxy, use this skill to trace the request through each provider, identify the specific failure point (e.g., incorrect URL, incompatible field, non-2xx status treated as success), and apply the correct fix.

Quick Start

Use the multi-provider-llm-proxy-debugging skill to investigate why the LLM proxy is returning empty responses for user queries.