skills.rest
jkenney47
jkenney47COMMUNITY
Explore all skills by this author →
GitHub

napkin

Log events, promote recurring patterns into rules, and rank memories for agent context.

Updated Aug 23, 2026
Search Tags:#context#ai agent#memory#persistence#rules#autopilot
One-click install
npx skills add https://github.com/jkenney47/finance-app-ux-planning-snapshot-20260308-155100 --skill napkin-jkenney47
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: napkin
Source: https://github.com/jkenney47/finance-app-ux-planning-snapshot-20260308-155100/tree/main/.agents/skills/napkin
Command: npx skills add https://github.com/jkenney47/finance-app-ux-planning-snapshot-20260308-155100 --skill napkin-jkenney47

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill addresses the problem of AI context loss and repeated mistakes by maintaining a persistent, structured working memory for the agent.

Core Features & Use Cases

  • Structured Event Log: Captures mistakes and corrections as events for future reference.
  • Rule Promotion: Identifies recurring patterns and promotes them into stable rules.
  • Memory Ranking: Prioritizes relevant memories to inject into the AI's context.
  • Use Case: When an AI agent repeatedly makes the same error, like misinterpreting a file path, this Skill logs the mistake and correction. Over time, it promotes this into a rule, ensuring the AI avoids the error in future sessions.

Quick Start

Use the napkin skill to start memory context for the current task.

Frequently Asked Questions about napkin

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

FAQPage Schema
How do I give an AI agent persistent memory across sessions?▼

Persistent memory for AI agents is maintained through a structured event log that captures mistakes and corrections, preventing context loss across sessions. A ranking model prioritizes relevant memories for injection into the current context.

How does rule promotion work for recurring AI agent mistakes?▼

Rule promotion identifies recurring patterns from the structured event log and promotes them into stable rules. This ensures the AI agent avoids repeated errors, such as misinterpreting file paths, in future task executions.

What is the best way to prevent AI agents from repeating the same errors?▼

The best way to prevent repeated errors is logging mistakes and corrections as structured events. Over time, recurring patterns are promoted into stable rules that automatically guide the AI agent away from those mistakes.

Do I need to set up a database to use persistent memory for AI agents?▼

No external database is required. The Skill manages persistent working memory entirely through a structured event log and rule curation, offering an advanced implementation without dependencies for immediate context retention.

How is relevant memory retrieved and injected into an AI agent's context?▼

Relevant memory is retrieved using a memory ranking model that prioritizes past events and rules. The highest-ranked memories are then injected into the AI agent's context to facilitate accurate task execution.

Related Skills

Explore More Software Engineering Skills→
cdrgurucdrguru
1

claude-code-init

Initialize Claude Code sessions with PACK context and session logs.

Community
Intermediate
ckallumckallum

prevent

Identify and implement deterministic prevention for Claude Code workflow mistakes and anti-patterns.

Community
Advanced
arbazkhan971arbazkhan971
26

agent

Design and validate autonomous AI agents with tool orchestration and guardrails.

Community
Advanced
DektoraDektora

rotation-handoff

Generate and read structured DekSpec session handoff records in dekspec/.scratch/rotation-handoff/.

Official
Advanced