Technical guides, research notes, and operator dispatches on AI agents, model infrastructure, and secure deployment.
CodeExecutor — AutoGen skill on mr.technology. Audit verdict: PENDING.
deep-research — Claude Skills Registry skill on mr.technology. Audit verdict: PENDING.
github-pr-creator — Claude Skills Registry skill on mr.technology. Audit verdict: PENDING.
000-jeremy-content-consistency-validator — Claude Skills Registry skill on mr.technology. Audit verdict: PENDING.
docker-manager — Claude Skills Registry skill on mr.technology. Audit verdict: PENDING.
langchain-azure-ai — LangChain skill on mr.technology. Audit verdict: PENDING.
self-improvement — LobeHub skill on mr.technology. Audit verdict: PENDING.
000-tnr — Claude Skills Registry skill on mr.technology. Audit verdict: PENDING.
langchain-openai — LangChain skill on mr.technology. Audit verdict: PENDING.
twitter-openclaw — LobeHub skill on mr.technology. Audit verdict: PENDING.
writing-skills — Skills.sh skill on mr.technology. Audit verdict: PENDING.
aaveclaw — clawhub skill on mr.technology. Audit verdict: PENDING.
001-polish-and-publish — Claude Skills Registry skill on mr.technology. Audit verdict: PENDING.
webapp-testing — LobeHub skill on mr.technology. Audit verdict: PENDING.
web-search — clawhub skill on mr.technology. Audit verdict: PENDING.
asana-pat — Claude Skills Registry skill on mr.technology. Audit verdict: PENDING.
004-exploration-and-review-skills — Claude Skills Registry skill on mr.technology. Audit verdict: PENDING.
aibtc — clawhub skill on mr.technology. Audit verdict: PENDING.
00-andruia-consultant — Claude Skills Registry skill on mr.technology. Audit verdict: PENDING.
00-basic-skill — Claude Skills Registry skill on mr.technology. Audit verdict: PENDING.
00-build-your-testing-skill — Claude Skills Registry skill on mr.technology. Audit verdict: PENDING.
Most browser automation falls apart the moment a shadow DOM shows up or a network request stalls. Playwright-Pro is the Architect-tier skill that treats your web app like a serious system — with structured page models, network interception, and headless execution that actually mirrors real user behavior.
Every AI agent degrades. Prompts drift, context windows fill with noise, and yesterday's effective strategy becomes today's hallucination fuel. The Self-Improving-Agent skill closes that loop — giving your Claude-Code agents the ability to observe their own failures, diagnose root causes, and patch their behavior before the next execution cycle.
Single-agent systems are straightforward. Multi-agent systems are a different beast — you need explicit role definitions, communication protocols, conflict resolution, and a clear theory of how emergent behavior arises from agent interactions. Agent-Designer forces you to answer those questions before you write a single line of agent code.