memory-organization

What problem does it solve?

Datus projects often accumulate messy persistent stores (semantic models, metrics, reference SQL, knowledge bases, agent memory, AGENTS.md files, and custom skills) with misclassified items, duplicate entries, conflicting facts, and stale or erroneous content that breaks agent accuracy and data consistency.

Core Features & Use Cases

• Full Store Inventory: Enumerates all contents of every persistent store in the project to build a complete baseline of existing data.
• Rule-Based Classification Audit: Cross-checks all items against the storage-classify skill's decision tree to spot misclassifications, duplicates, conflicts, and stale entries.
• Safe Remediation Workflow: Generates a clear, user-confirmable remediation plan before executing any fixes, with guardrails to prevent destructive changes without approval.
• Use Case: A data team using Datus can run this skill to clean up their knowledge base, removing duplicate metric definitions and misclassified SQL snippets to ensure the agent generates accurate, consistent responses for analysts.

Quick Start

Invoke the memory-organization skill to audit all persistent data stores in your Datus project and receive a remediation plan for any classification, duplication, or consistency issues detected.