VAMFIOfficial·5 Agent Skills Included

claude-user-memory

Research-first coding workflows with persistent memory and quality gates

Enforces a research, plan, implement workflow for coding tasks so nothing is built from stale or guessed API knowledge. Adds persistent memory, pattern capture, and context curation so past decisions and learnings carry across sessions. Blocks low-quality work with scoring gates, rollback plans, and automatic test verification before code ships.

206|32

↗

npx skills add VAMFI/claude-user-memory --all -g -y

Available:

Instructs the agent to route every coding request through the research, plan, implement workflow, use the nine specialist agents and five auto-invoked skills, and apply extended thinking levels for complex decisions.

All Skills in This Repository (5)

Pure Emerald Level Indicators

📦 In Repo

VAMFI

context-engineering

Automates workspace cleanup and improves RAG performance by selecting what to forget and what to remember.

research-methodology

Automate authoritative API documentation research with version detection and explicit citations.

quality-validation

Validate ResearchPacks and Implementation Plans with predefined rubrics and scoring.

pattern-recognition

Extract reusable code patterns from completed implementations and index them for reuse.

planning-methodology

Create reversible implementation plans with risk assessment and verification steps.

Official

Intermediate

Frequently Asked Questions

FAQPage Schema

How to install claude-user-memory?▼

Run `npx skills add VAMFI/claude-user-memory --all -g -y` in your terminal to install all skills globally.

How to stop AI coding from outdated API knowledge?▼

This system forces a research phase that fetches version-accurate official documentation before any code is written, cutting hallucinated API errors dramatically.

Does it remember decisions between sessions?▼

Yes. A persistent knowledge core captures patterns, decisions, and learnings after each task and reloads them in future sessions.

What are the quality gates in this workflow?▼

Research outputs must score 80+ and plans 85+ before coding begins, and implementations must pass tests with a three-retry circuit breaker.

Can non-developers use this workflow?▼

Yes. You describe the feature in plain English and the agents handle research, planning, implementation, and verification automatically.