Does this skill make changes automatically?

No. The skill identifies potential improvements and presents them to the user for validation and approval before making any file updates.

When does the reflection process start?

It triggers after an agent believes it has satisfied a user request but before providing the final response, allowing it to capture context before the session ends.

What is the Meta-Reflection skill?

It is a specialized capability for Claude Code that triggers a post-task review to identify ways to improve agents, tools, and configurations based on the session's performance.

Can I skip reflection for simple tasks?

Yes, the protocol is designed to skip reflection for trivial requests, quick lookups, or when the user explicitly requests no review.

Which files does this skill typically update?

It primarily recommends updates to AGENTS.md, CLAUDE.md, .instructions.md, and individual SKILL.md files to refine agent behavior and context.

Meta-Reflection & Continuous Improvement

Name: Meta-Reflection & Continuous Improvement
Author: doc-detective

bydoc-detective

0•

Productivity & Workflow

Analyzes completed tasks to identify and implement improvements for agents, tools, and project workflows.

The Meta-Reflection skill empowers Claude Code agents to learn from every interaction by conducting a systematic review after task completion. By analyzing friction points, backtracking, and missed opportunities during a session, the skill identifies actionable updates for AGENTS.md, CLAUDE.md, and individual skill configurations. It transforms one-off troubleshooting into permanent system optimizations, ensuring that agents become more efficient, accurate, and context-aware with every subsequent request.

Key Features

01Actionable recommendations for AGENTS.md and CLAUDE.md

02Seamless implementation of approved system improvements

03Systematic session review and friction analysis

040 GitHub stars

05Structured presentation of findings for user approval

06Automated identification of missing skill triggers and context

Use Cases

01Reducing repetitive manual workflows by proposing new skill patterns

02Optimizing agent behavior after complex debugging sessions

03Identifying and filling documentation gaps in project-specific instructions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add doc-detective/agent-tools reflection

For use in Claude.ai and ChatGPT

Download Skill