mci

Capture single transferable insights with trigger context for agent learning.

8|Updated Feb 2, 2026
One-click install
npx skills add https://github.com/usurobor/cnos --skill mci
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mci
Source: https://github.com/usurobor/cnos/tree/main/src/packages/cnos.core/skills/agent/mci
Command: npx skills add https://github.com/usurobor/cnos --skill mci

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Identify and capture the minimum coherent insights that prevent repeating the same mistakes in future agent cycles.

Core Features & Use Cases

  • Explicit capture of single, transferable insights with trigger context.
  • Migration-ready: designed for transfer into higher-level skills and long-term memory.
  • On-demand activation: loads when needed during task execution and model updates.

Quick Start

Instruct the agent to extract the smallest lesson that changes its next action and record it with its trigger.

Frequently Asked Questions about mci

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

FAQPage Schema
How do I stop my AI agent from repeating the same mistakes in future cycles?

To prevent repeating mistakes, you capture the minimum coherent insight that changes the agent's next action. This skill extracts single, transferable lessons with explicit trigger context during task execution, ensuring errors do not recur in subsequent agent cycles.

What is the best way to capture task-local learning insights during agent execution?

The best way to capture task-local learning is to instruct the agent to extract the smallest lesson that changes its next action. This requires explicit trigger capture, imperative phrasing, and single insights per capture to ensure safe, auditable learning within the agent cycle.

Can I migrate captured insights into higher-level skills for long-term memory?

Yes, you can migrate captured insights into higher-level skills and long-term memory. The captured insights are migration-ready, designed with explicit transfer pathways to move task-local learning into the broader skill framework for future agent cycles.

Does on-demand insight capture work without adding extra dependencies?

Yes, on-demand insight capture works without extra dependencies. The skill loads only when needed during task execution and model updates, requiring no external components to identify and record transferable insights with their trigger context.

What guardrails are needed for safe and auditable agent learning?

Safe and auditable agent learning requires guardrails like explicit trigger capture, imperative phrasing, and single insights per capture. These constraints ensure that migration pathways into higher-level skills remain controlled and auditable during on-demand task execution.