Can I preview changes before they are applied?

Yes, by setting the dry_run parameter to true, you can generate a detailed report of exactly what would be updated, pruned, or extracted without making any permanent changes.

What is the difference between light and full consolidation?

Light consolidation is a fast daily maintenance process that updates stability scores and prunes weak links. Full consolidation is a compute-intensive process that uses ML clustering for deep knowledge reorganization and entity extraction.

Does consolidation delete my important memories?

No, it optimizes them. It only prunes 'orphaned' or extremely weak relationships that create noise. Core episode content remains intact while the connections between them are refined.

What happens if I don't have the HDBSCAN library installed?

If full consolidation is requested without HDBSCAN, the skill gracefully falls back to light consolidation, logs a warning, and suggests the necessary installation command.

How often should I run memory consolidation?

Light consolidation is recommended daily or after 20+ new episodes. Full consolidation is best performed weekly or after 100+ episodes have been recorded.

World Weaver Memory Consolidator

Name: World Weaver Memory Consolidator
Author: astoreyai

byastoreyai

•

Gestión de Bases de Datos

Optimizes AI memory structures through automated clustering, pruning, and relationship strengthening to maintain long-term context efficiency.

Acerca de

The World Weaver Memory Consolidator skill implements a sophisticated memory maintenance system for Claude, mimicking biological memory processes to handle large-scale information retrieval. It optimizes long-term storage by clustering similar episodes into themes, applying FSRS-based stability decay to stale data, and strengthening frequently accessed connections. By performing either 'Light' daily maintenance or 'Full' deep restructuring using HDBSCAN clustering and entity extraction, this skill ensures that Claude's contextual memory remains fast, relevant, and well-organized even as project complexity grows.

Características Principales

1 GitHub stars
Automated entity extraction and duplicate merging in knowledge graphs
Detailed health reporting with dry-run capabilities for previewing changes
Advanced HDBSCAN clustering for thematic grouping of memory episodes
Dual-mode consolidation: Fast 'Light' updates or intensive 'Full' restructuring
FSRS-based stability decay to prioritize frequently accessed information

Casos de Uso

Optimizing slow memory search performance after accumulating a large volume of project interactions
Archiving and organizing domain-specific knowledge after reaching major development milestones
Maintaining a lean and efficient long-term AI context through scheduled memory maintenance

Acerca de

Características Principales

1 GitHub stars
Automated entity extraction and duplicate merging in knowledge graphs
Detailed health reporting with dry-run capabilities for previewing changes
Advanced HDBSCAN clustering for thematic grouping of memory episodes
Dual-mode consolidation: Fast 'Light' updates or intensive 'Full' restructuring
FSRS-based stability decay to prioritize frequently accessed information

Casos de Uso

Optimizing slow memory search performance after accumulating a large volume of project interactions
Archiving and organizing domain-specific knowledge after reaching major development milestones
Maintaining a lean and efficient long-term AI context through scheduled memory maintenance