Can I use Gladiator to track architectural decisions?

Yes, it includes specific templates for recording domain conventions and architectural decisions, making it a valuable tool for project documentation.

What is the primary purpose of Claude Gladiator?

It creates a feedback loop where Claude learns from its mistakes and your corrections, eventually updating its own rules to provide better performance in your specific project.

How does the reflection process work?

The gladiator_reflect() command analyzes your stored observations, clusters them into logical recommendations, and proposes specific updates to your existing project artifacts.

Does Gladiator record my data automatically?

No, while it provides nudges after tool failures, you must manually trigger the observation or reflection commands to ensure only accurate and relevant information is captured.

How do I start a batch review of all my observations?

You can use the /review-gladiator command to process all accumulated observations from your session history and generate global improvement recommendations.

Claude Gladiator

Name: Claude Gladiator
Author: Vvkmnn

byVvkmnn

0•

分析与监控

Implements a continuous learning loop that observes tool failures and user corrections to evolve AI rules and behaviors.

Claude Gladiator is a powerful self-improvement engine for Claude Code that transforms errors into institutional knowledge. By monitoring tool failures and manual corrections, the skill prompts the user to record specific patterns and successful fixes, preventing the AI from making the same mistake twice. It systematically clusters observations into actionable recommendations, allowing you to update your project's rules, hooks, and skills based on real-world usage. This ensures that your AI coding assistant becomes increasingly specialized and efficient the more you use it.

主要功能

01Context-aware observation logging for tool failures and user corrections

02Batch review functionality to consolidate learning across multiple sessions

03Automated nudges for reflection after Bash or Edit tool errors

04Intelligent clustering of session history into actionable rule updates

05Seamless integration with existing project artifacts to prevent duplication

060 GitHub stars

使用场景

01Scaling AI proficiency across complex codebases by refining environment-specific rules

02Eliminating repetitive tool misuse by documenting and applying successful fix patterns

03Standardizing project-specific conventions and architectural decisions as they emerge

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add vvkmnn/claude-emporium-beta claude-gladiator

For use in Claude.ai and ChatGPT

主要功能

01Context-aware observation logging for tool failures and user corrections

02Batch review functionality to consolidate learning across multiple sessions

03Automated nudges for reflection after Bash or Edit tool errors

04Intelligent clustering of session history into actionable rule updates

05Seamless integration with existing project artifacts to prevent duplication

060 GitHub stars

使用场景

01Scaling AI proficiency across complex codebases by refining environment-specific rules

02Eliminating repetitive tool misuse by documenting and applying successful fix patterns

03Standardizing project-specific conventions and architectural decisions as they emerge

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add vvkmnn/claude-emporium-beta claude-gladiator

For use in Claude.ai and ChatGPT