Can I move a project-specific tip to my global settings?

Yes, you can use the 'promote' command within the layered_merge script to move high-quality insights from the project layer to your global layer for use in other repositories.

Will updating my skills delete my saved experiences?

No. Because experiences are stored in separate evolution.json files, the 'align_all' script can re-stitch your saved knowledge back into the SKILL.md files even after a plugin update.

How do I trigger the evolution process?

Simply ask Claude to 'evolve', 'review the conversation', or 'save this experience' at the end of a task to trigger the diagnostic and storage workflow.

What is the three-layer architecture?

It organizes experience into three levels: Upstream (community defaults), Global (your personal preferences across all projects), and Project (specific to a single repository), allowing for tailored behavior at every level.

Skill Evolution Manager

Name: Skill Evolution Manager
Author: echoleesong

byecholeesong

Productivity & Workflow

Captures and persists AI conversation insights into a reusable knowledge base across projects and versions.

About

Skill Evolution Manager acts as the 'evolutionary hub' for the Claude Code skill ecosystem, transforming fleeting chat insights into permanent, structured best practices. It solves the common problem of knowledge loss during skill updates by utilizing a unique three-layer architecture—Upstream, Global, and Project—that separates communal, personal, and project-specific experiences. By intelligently stitching these insights back into skill documentation, it ensures that your AI assistant grows smarter and more aligned with your specific coding style and environment over time.

Key Features

0 GitHub stars
Automated 'Post-Mortem' diagnosis to identify successful patterns and error fixes
Structured JSON extraction of user preferences, context, and custom prompts
Smart stitching technology that merges experiences directly into SKILL.md files
Cross-version alignment to recover custom knowledge after plugin updates
Three-layer experience management (Upstream, Global, Project) for hierarchical knowledge

Use Cases

Preserving specific environment configurations and bug workarounds across sessions
Synchronizing personal coding style preferences across multiple development projects
Extracting high-quality architectural patterns from conversations to create community PRs

About

Key Features

0 GitHub stars
Automated 'Post-Mortem' diagnosis to identify successful patterns and error fixes
Structured JSON extraction of user preferences, context, and custom prompts
Smart stitching technology that merges experiences directly into SKILL.md files
Cross-version alignment to recover custom knowledge after plugin updates
Three-layer experience management (Upstream, Global, Project) for hierarchical knowledge

Use Cases

Preserving specific environment configurations and bug workarounds across sessions
Synchronizing personal coding style preferences across multiple development projects
Extracting high-quality architectural patterns from conversations to create community PRs