context-maintenance

Diagnose and fix drift between memory sources and the codebase.

1|2|Updated Apr 6, 2026
One-click install
npx skills add https://github.com/parisgroup-ai/imersao-ia-setup --skill context-maintenance-parisgroup-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: context-maintenance
Source: https://github.com/parisgroup-ai/imersao-ia-setup/tree/main/skills/context-maintenance
Command: npx skills add https://github.com/parisgroup-ai/imersao-ia-setup --skill context-maintenance-parisgroup-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Audits and reconciles Claude context sources (memory files, CLAUDE.md, AGENTS.md) against the live codebase to detect and fix stale or inconsistent information, preventing outdated context from steering conversations.

Core Features & Use Cases

  • Memory index health checks and drift detection between MEMORY.md and codebase.
  • AGENTS.md completeness and cross-check against actual services and files.
  • Stale topic file detection and recommended maintenance to keep memory aligned with the project.
  • Timestamp consistency and structural integrity checks across CLAUDE.md, AGENTS.md, and memory topics.

Quick Start

Run a full context audit to identify drift between memory sources and the codebase, then apply fixes to memory and AGENTS.md to restore alignment.

Frequently Asked Questions about context-maintenance

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I fix Claude context drift between memory files and my codebase?

To fix Claude context drift, audit memory sources like MEMORY.md and AGENTS.md against the live codebase to detect stale information, then apply targeted updates to restore alignment and prevent outdated context from steering conversations.

Why does Claude have outdated context from previous sessions?

Claude has outdated context when memory index files, AGENTS.md, or topic files fall out of sync with the current codebase state, causing structural drift and stale timestamps that misdirect AI reasoning during new sessions.

How do I run a full context audit for projects using CLAUDE memory?

Run a full context audit by checking memory index health, verifying AGENTS.md completeness against actual services, detecting stale topic files, and validating timestamp consistency across CLAUDE.md to identify and fix structural inconsistencies.

What is the best way to maintain AGENTS.md completeness for Claude?

The best way to maintain AGENTS.md completeness is cross-checking it against actual services and files in the codebase, enforcing consistency by updating it alongside MEMORY.md and topic files to accurately reflect the current project state.

Do I need CLAUDE.md to detect stale memory topic files?

You need CLAUDE.md and related memory files to detect stale topic files effectively, as the audit process checks timestamp consistency and structural integrity across CLAUDE.md, AGENTS.md, and memory topics to ensure accurate context alignment.