Zikra — AI Persistent Memory

Name: Zikra — AI Persistent Memory
Availability: InStock
Author: getzikra

Store, search, and log project memories across AI sessions via ZIKRA_URL.

8|1|Updated Mar 30, 2026

Search Tags:#error logging #project context #hybrid search #decision tracking #ai persistent memory #prompt reuse #self-hosted mcp

Ready to run

npx skills add https://github.com/getzikra/Zikra --skill zikra-ai-persistent-memory

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: Zikra — AI Persistent Memory
Source: https://github.com/getzikra/Zikra/tree/main
Command: npx skills add https://github.com/getzikra/Zikra --skill zikra-ai-persistent-memory

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Zikra solves the problem of losing important project context between AI sessions and across machines by providing a governed, cross-agent persistent memory store.

Core Features & Use Cases

Search and retrieve relevant past decisions, errors, prompts, and context at the start of a session.
Save structured memories immediately after key decisions so the right context stays available for later runs.
Log failures by recording errors with supporting context so they can be tracked and found again.
Use typed memory categories such as decision, conversation, error, schema, and prompt to keep retrieval precise.

Quick Start

Tell your AI agent to store a new decision memory for project yourproject with title and content, then immediately search for related context to continue working.

Frequently Asked Questions about Zikra — AI Persistent Memory

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

FAQPage Schema

How do I save AI project context and decisions across sessions?▼

AI persistent memory stores decisions, errors, and prompts in scoped project namespaces so teams can retrieve them later. It applies at session start for retrieval, after key decisions for saving, and on failures for error logging.

What is the best way to track AI errors and decisions across multiple agents?▼

The best way to track AI errors and decisions across agents is using a governed persistent memory store with typed categories like decision, conversation, error, schema, and prompt to keep retrieval precise across machines.

How do I retrieve past prompts and project context at the start of a new AI session?▼

You retrieve past prompts at session start by searching scoped project namespaces using a persistent memory tool. This searches past decisions, errors, and context so the right information stays available for continuing work.

Do I need a self-hosted MCP to enable cross-agent persistent memory?▼

A self-hosted MCP is used to enable secure cross-agent persistent memory. It satisfies a YAML-frontmatter-defined skill interface and executes commands securely through configured URL and TOKEN environment variables for project namespaces.

Can I categorize AI memory types to keep retrieval precise for complex projects?▼

Yes, you can use typed memory categories such as decision, conversation, error, schema, and prompt to keep retrieval precise. This structured approach ensures relevant context stays available for later runs.

How do I log AI agent failures with supporting context for later debugging?▼

You log AI agent failures by recording errors with supporting context on failure events using a persistent memory tool. This captures the error within scoped project namespaces so it can be tracked and found again.