skills.rest
trungtnm
trungtnmCOMMUNITY
Explore all skills by this author →
GitHub

cm

Persist coding agent knowledge across sessions with a Bun/TypeScript CLI.

1|Updated Feb 3, 2026
Search Tags:#ai#learning#patterns#memory#coding#cross-agent
One-click install
npx skills add https://github.com/trungtnm/claude-code-utils --skill cm-trungtnm
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: cm
Source: https://github.com/trungtnm/claude-code-utils/tree/main/plugins/skills/cm
Command: npx skills add https://github.com/trungtnm/claude-code-utils --skill cm-trungtnm

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Procedural memory for AI coding agents to persist knowledge across sessions and agents, turning scattered logs into actionable guidance and shared learning.

Core Features & Use Cases

  • Cross-agent memory: share patterns, decisions, and lessons across Claude Code, Cursor, Codex, and other agents.
  • Three-layer cognitive architecture: episodic memory, working memory, and procedural memory with confidence decay and safety guardrails.
  • Safety and governance: trauma guard safeguards, anti-pattern learning, and cross-project memory for consistent coding standards.

Quick Start

Run the CM context command to extract applicable procedural-memory rules for your current coding task.

Frequently Asked Questions about cm

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

FAQPage Schema
How do I share memory and learning patterns across multiple AI coding agents?▼

You can share memory across AI coding agents by using a cross-agent memory system that captures patterns, decisions, and lessons from tools like Claude Code and Cursor. This transforms scattered session logs into persistent procedural memory for consistent development.

What is procedural memory for AI coding assistants and when do I need it?▼

Procedural memory for AI coding assistants is a persistent knowledge base of patterns and decisions that survives across sessions. You need it when agents repeatedly lose context, forcing you to re-explain project standards and past architectural decisions.

Can I use cross-agent memory with different coding tools like Claude Code and Cursor?▼

Yes, cross-agent memory works with different coding tools like Claude Code, Cursor, and Codex. It extracts applicable procedural-memory rules via a Bun/TypeScript CLI, ensuring shared learning and consistent coding standards across your entire AI development workflow.

How do I extract applicable procedural memory rules for a current coding task?▼

You extract applicable procedural-memory rules by running the CM context command via its Bun/TypeScript CLI. This fetches relevant patterns and past decisions, providing actionable guidance directly into your current development task.

What are the limitations of relying on session logs for AI coding agent memory?▼

Session logs lack persistence and cross-agent context, meaning patterns and decisions are lost when sessions end. Without a dedicated procedural memory architecture, agents cannot share learning safely across different projects or prevent repeating anti-patterns.

Related Skills

Explore More Software Engineering Skills→
DicklesworthstoneDicklesworthstone
70

cm

Consolidate coding sessions into a persistent cross-agent memory store.

Community
Advanced
gaoqiongxiegaoqiongxie
1

beads-memory

Capture codebase snapshots and load compressed context across sessions.

Community
Intermediate
akillnessakillness
40

react-grab

Capture React component names, file paths, and HTML from browser UI elements.

Community
Advanced
JeffreyyvdbJeffreyyvdb
1

dotnet-inspect

Query, compare, and explore .NET APIs from NuGet packages and local files.

Community
Advanced