agent-introspection-debugging

Diagnose and recover failing agent runs through structured introspection.

86|21|Updated Feb 9, 2026

Search Tags:#debugging #agent #diagnosis #introspection #recovery #failure

Ready to run

npx skills add https://github.com/Jamkris/everything-gemini-code --skill agent-introspection-debugging-jamkris

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agent-introspection-debugging
Source: https://github.com/Jamkris/everything-gemini-code/tree/main/skills/agent-introspection-debugging
Command: npx skills add https://github.com/Jamkris/everything-gemini-code --skill agent-introspection-debugging-jamkris

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a structured self-debug workflow for AI agents that fail repeatedly, loop on tools, or drift from the task, enabling systematic failure capture, diagnosis, containment recovery, and a clear introspection report.

Core Features & Use Cases

Four-Phase Loop: Failure Capture, Root-Cause Diagnosis, Contained Recovery, and Introspection Report to guide reliable debugging before escalation.
ECC Integration: Aligns recovery with verification loops and continuous-learning processes to improve long-term agent reliability.
Output & Artifacts: Produces a human-readable debug report documenting failure, diagnosis, recovery actions, and results for audit and learning.

Quick Start

Initiate the introspection workflow on a failing agent run to capture failure details, diagnose root causes, apply contained recovery, and generate a self-debug report.

Frequently Asked Questions about agent-introspection-debugging

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

FAQPage Schema

How do I debug an AI agent that is stuck in a tool calling loop?▼

To debug an agent stuck in a tool calling loop, you can apply a structured introspection workflow that captures the failure, diagnoses the root cause, applies contained recovery, and generates an audit report.

What is structured introspection for recovering from failing agent runs?▼

Structured introspection for failing agent runs is a deterministic debugging process divided into four phases: failure capture, root-cause diagnosis, contained recovery, and an introspection report to ensure safe and auditable recovery.

How do I diagnose an autonomous agent drifting from its original goals?▼

You can diagnose an agent drifting from its goals by initiating an introspection workflow that captures the failure details and applies root-cause diagnosis before attempting contained recovery actions.

Does agent introspection integrate with continuous-learning processes for safe recovery?▼

Yes, agent introspection integrates with ECC to align recovery actions with verification loops and continuous-learning processes, improving long-term agent reliability while ensuring safe and auditable recovery.

What artifacts are produced when diagnosing agent failures with introspection?▼

Diagnosing agent failures with introspection produces a human-readable debug report that documents the original failure, root-cause diagnosis, recovery actions taken, and final results for audit and learning purposes.

When should I use a structured self-debug workflow for AI agents?▼

You should use a structured self-debug workflow when your AI agents fail repeatedly, loop on tools, or drift from the task, enabling systematic failure capture and contained recovery before escalation.