Learner Skill

功能

• Extracts learned skills from conversations
• Defines expertise criteria for quality skills
• Provides a structured workflow for skill extraction
• Classifies learnings into Expertise or Workflow types
• Saves skills in project- or user-level locations

使用场景

• When you solve a tricky bug and want to document the principle for future reference
• When you discover a non-obvious workaround specific to your codebase
• When you uncover undocumented behavior that impacts the project
• To create a repository of team-specific problem-solving heuristics

非目标

• Extracting generic programming patterns that are easily Googleable
• Documenting standard refactoring techniques
• Saving simple library usage examples
• Capturing basic type definitions or boilerplate code

工作流

1. Gather specific error, symptom, or confusion, including exact error messages, file paths, and line numbers.
2. Document the exact fix, including code snippets, file paths, and configuration changes.
3. Identify keywords (error fragments, file names, symptoms) that signal the problem for future triggers.
4. Validate the skill against quality criteria: specificity, non-Googleability, precise solutions, and good triggers.
5. Classify the learned insight as either 'Expertise' (domain knowledge, pattern, gotcha) or 'Workflow' (operational procedure).
6. Save the skill to the appropriate location (project-level or user-level) using the required YAML frontmatter format.

实践

• Skill extraction methodology
• Knowledge formalization
• Pattern recognition

先决条件

• Claude Code CLI
• Claude Max/Pro subscription or Anthropic API key