kernel-debug

Map agent reasoning failure symptoms to kernel phases and source files.

21|4|Updated Feb 19, 2026
One-click install
npx skills add https://github.com/tylerjrbuell/reactive-agents-ts --skill kernel-debug-tylerjrbuell
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: kernel-debug
Source: https://github.com/tylerjrbuell/reactive-agents-ts/tree/main/.agents/skills/kernel-debug
Command: npx skills add https://github.com/tylerjrbuell/reactive-agents-ts --skill kernel-debug-tylerjrbuell

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps engineers rapidly diagnose agent reasoning failures by mapping observed symptoms to the exact kernel phases and source files to inspect, avoiding broad codebase exploration and guesswork.

Core Features & Use Cases

  • Symptom → Phase Mapping: A concise table that maps common failure symptoms (missing tool calls, loops, silent failures, context compaction, final-answer gating) to the responsible kernel phase and files.
  • Targeted Grep Commands: Ready-to-run grep commands tailored to each symptom to quickly find related code and configuration.
  • Debug Workflow & Tips: Practical steps including enabling full model I/O logging and choosing whether to inspect state.messages[] or state.steps[] for root-cause analysis.
  • Use Case: When an agent never calls tools, use the mapping to check context-builder, guard chains, and FC schema generation instead of searching the entire repository.

Quick Start

Use the kernel-debug guidance to map an observed symptom such as agent never calls tools to the specific kernel phase, files, and grep commands to inspect.

Frequently Asked Questions about kernel-debug

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

FAQPage Schema
How do I debug an agent that never calls tools in a TypeScript Reactive Agents codebase?

To debug missing tool calls in TypeScript Reactive Agents, map the symptom to the context-builder, guard chains, and function-calling schema generation phases. Use targeted grep commands to inspect those specific source files instead of searching the entire repository.

What is the best way to diagnose silent LLM failures during agent reasoning?

Diagnosing silent LLM failures involves mapping the symptom to specific kernel phases like final-answer gating or context compaction. Enable full model I/O logging and inspect state.messages[] or state.steps[] to perform rapid root-cause analysis without executing code.

How do I fix infinite thought loops in TypeScript Reactive Agents?

Fixing infinite thought loops requires mapping the agent failure to the responsible kernel phase and source files. Apply targeted grep commands to quickly locate related code and configuration, then inspect state steps to identify the root cause of the looping behavior.

Can I use grep commands to find the source of agent reasoning failures without executing code?

Yes, you can use provided grep commands to find the source of agent reasoning failures without executing code. The mapping table supplies ready-to-run grep commands tailored to each symptom to quickly locate related code and configuration for root-cause analysis.

When should I inspect state.messages[] instead of state.steps[] for root-cause analysis?

Choosing between inspecting state.messages[] and state.steps[] depends on the specific agent reasoning failure observed. The debug workflow provides practical tips to help you decide which state array to inspect for rapid root-cause analysis of symptoms like context compaction or missing tool calls.