Audit and repair filesystem-based AI memory systems.
npx skills add https://github.com/zircote/mnemonic --skill custodian
Please help me install this Agent Skill. Skill: custodian Source: https://github.com/zircote/mnemonic/tree/main/skills/custodian Command: npx skills add https://github.com/zircote/mnemonic --skill custodian
This Skill ensures the reliability and accuracy of your AI's memory system by performing regular health checks, validating data integrity, and fixing broken links or inconsistencies.
custodian audit --fix command to ensure all internal references are updated and no memories have become orphaned.Run a full audit of the memory system and fix any detected issues.
High-intent search queries and answers about installing and using this skill.
To validate AI memory system integrity, run an audit that checks frontmatter consistency, verifies internal links, and identifies orphaned memories. This process repairs broken references and updates decay scores to maintain memory health.
Memory decay management updates decay scores within the filesystem-based AI memory system to ensure ongoing relevance. By recalculating these scores, the system identifies outdated memories and maintains the accuracy of the knowledge base.
Yes, you can automatically fix orphaned memories after a knowledge base refactor. The system identifies memories lacking incoming references and optionally re-links them, ensuring all data remains connected and accessible.
The best way to maintain AI memory health is to perform regular audits that validate file integrity and repair internal links. Executing an automated audit with fix commands resolves broken links and manages decay scores.
Yes, Python is required to validate memory file integrity and frontmatter. The system operates using Python modules for parsing, validation, and file operations to ensure data consistency and reliability across the memory system.
Happyverse-TeamStore and retrieve project memories in .agent/memories.md via Ralph CLI commands.
fefogarciaDesign persistent memory architectures with vector stores and temporal knowledge graphs.
ConductionNLInitialize capture-driven documentation scaffolds with Playwright specs and Docusaurus configs.
Hidden-HistorySearch AI memory collections by semantic similarity with filtering and decay scoring.