What is the primary goal of the oe-trace-and-fallback-triage skill?

The primary goal is to help developers identify and fix instances where the AI returns a generic 'fallback' message instead of a specific, helpful response.

Is it safe to run these diagnostics in a production environment?

No, the skill specifies that diagnostic modes like OE_FALLBACK_DIAGNOSTICS should only be enabled in development (dev-only) to avoid performance impacts and log verbosity.

How does the skill help pinpoint the source of a failure?

It uses diagnostic flags to provide a 'Source' and 'Trigger' reason, which maps directly to specific Python files in the backend workflows, such as the verbalizer or extraction logic.

What kind of reproduction scripts does this skill support?

It supports backend-only smoke tests for intake routing and specific UX scenario scripts (like manual_ux_scenario_I.py) for testing site-visit interactions.

OpenEvent Trace & Fallback Triage

Name: OpenEvent Trace & Fallback Triage
Author: shami-ah

byshami-ah

分析与监控

Debugs and eliminates generic fallback responses by pinpointing failure triggers and automating reproduction steps.

关于

The oe-trace-and-fallback-triage skill is a specialized diagnostic tool for developers working with OpenEvent-AI to resolve 'generic stub' replies such as 'no specific information available.' It streamlines the debugging process by enabling structured diagnostic logs, activating execution traces, and providing a direct mapping between fallback symptoms and their underlying backend modules. Whether dealing with Q&A verbalization errors or complex event reuse logic bugs, this skill provides deterministic reproduction scripts and regression testing patterns to ensure assistant reliability.

主要功能

Deterministic reproduction scripts for intake and UX scenarios
Structured capture of client-facing symptoms and session context
Integrated regression testing against known 'Chain of Failure' traps
Automated diagnostic logging and trace activation via environment variables
0 GitHub stars
Source-to-module mapping for Q&A and LLM adapter fallbacks

使用场景

Identifying why intake routing is failing in backend workflows
Debugging empty or generic 'Thanks for your message' assistant responses
Resolving site-visit logic errors in complex multi-step event scenarios

关于

主要功能

Deterministic reproduction scripts for intake and UX scenarios
Structured capture of client-facing symptoms and session context
Integrated regression testing against known 'Chain of Failure' traps
Automated diagnostic logging and trace activation via environment variables
0 GitHub stars
Source-to-module mapping for Q&A and LLM adapter fallbacks

使用场景

Identifying why intake routing is failing in backend workflows
Debugging empty or generic 'Thanks for your message' assistant responses
Resolving site-visit logic errors in complex multi-step event scenarios