What files does the Knowledge Extractor update?

The skill primarily suggests updates to DISCOVERIES.md for capturing specific learnings and PATTERNS.md for reusable solutions, as well as suggesting new agent definitions in the .claude/agents/ directory.

How does this prevent 'session amnesia' in AI development?

It captures the ephemeral context and 'why' behind decisions from your conversation history and commits it to persistent Markdown files within your repository, making that knowledge available to future AI sessions and human developers.

When is the best time to use this skill?

It is most effective when used at the end of a work session, immediately after resolving a difficult bug, or after implementing a complex new feature.

Does it automatically create new agents without supervision?

No, it identifies high-value opportunities for automation based on your repeated workflows and provides a structured recommendation that you can review before implementation.

Knowledge Extractor

Name: Knowledge Extractor
Author: rysweet

byrysweet

•

Productivity & Workflow

Synthesizes architectural patterns, debugging insights, and workflow optimizations from coding sessions to build persistent project memory.

The Knowledge Extractor skill transforms ephemeral coding conversations and debugging sessions into a structured organizational knowledge base. By automatically analyzing successful solutions and failed attempts, it updates project documentation like DISCOVERIES.md and PATTERNS.md, ensuring that hard-won insights are preserved long after a session ends. This skill also identifies repetitive workflows ripe for automation, providing structured blueprints for new specialized agents to further streamline the development process and prevent the repetition of past mistakes.

Key Features

01Standardized updates to DISCOVERIES.md for capturing novel project insights

02Automated extraction of root causes and solutions from complex debugging sessions

03Identification and formalization of reusable architectural patterns in PATTERNS.md

04Cross-session knowledge preservation to ensure continuous team and system learning

05Intelligent triggers for recommending new specialized agents based on repeated workflows

0616 GitHub stars

Use Cases

01Conducting post-mortems after resolving complex production bugs or configuration issues

02Formalizing ad-hoc architectural decisions into a reusable organizational design library

03Scaling development teams by building a searchable, shared repository of project-specific tribal knowledge

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add rysweet/amplihack knowledge-extractor

For use in Claude.ai and ChatGPT

Download Skill