debug-inference

Diagnose OpenShell inference path failures across gateways, providers, and network policies.

Updated Apr 20, 2026
One-click install
npx skills add https://github.com/x402agent/openclawd-stack --skill debug-inference
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: debug-inference
Source: https://github.com/x402agent/openclawd-stack/tree/main/openclawd-stack/OpenShell-main/.agents/skills/debug-inference
Command: npx skills add https://github.com/x402agent/openclawd-stack --skill debug-inference

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Diagnose why OpenShell inference is failing (managed inference and external inference) by identifying root causes across gateways, providers, and network policies, enabling fast, reliable repairs.

Core Features & Use Cases

  • Auto-diagnostics of managed vs external inference routes using OpenShell CLI primitives such as status, gateway info, and provider config checks
  • Sandbox-based probing to verify end-to-end connectivity (curl inside a sandbox) and verify common inference endpoints
  • Step-by-step workflow guidance to isolate gateway topology, configuration, and provider mismatches
  • Repair guidance to reconfigure inference routes and validate results post-fix

Quick Start

Run the OpenShell inference diagnostic workflow by executing status, inference get, provider get, and a sandbox probe to confirm fixes.

Frequently Asked Questions about debug-inference

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
Why is my OpenShell inference failing across different gateways and providers?

OpenShell inference fails when gateway topology, provider configurations, or network policies are mismatched. Diagnosing these routes requires checking gateway status, inspecting provider settings, and running sandbox probes to isolate the root cause.

How do I diagnose OpenShell inference routing issues step by step?

Diagnose OpenShell inference routing by running CLI status checks, retrieving inference and provider configurations, and executing sandbox probes to verify end-to-end connectivity. This workflow isolates failures across managed and external routes.

Can I test inference endpoint connectivity from inside a sandbox environment?

You can verify inference endpoint connectivity by running curl commands inside a sandbox environment. This sandbox-based probing tests end-to-end network reachability and validates that inference routes are correctly configured.

What is the best way to repair broken external inference routes in OpenShell?

The best way to repair broken external inference routes is to follow structured diagnostic workflows that identify provider mismatches, reconfigure route settings, and validate results post-fix using sandbox probes.

Does this diagnostic workflow cover both managed and external inference paths?

The diagnostic workflow covers both managed and external inference paths. It auto-diagnoses routing failures by comparing gateway info and provider configs, then validates repairs across both inference route types.

OpenShell inference is still not working after reconfiguring providers, what should I check next?

When OpenShell inference still fails after provider reconfiguration, check gateway topology status and run sandbox probes to verify network policies. Post-fix validation ensures the reconfigured inference routes are reachable.