mberto-compound

Core infrastructure with standard MCP servers (Linear, Context7, Langfuse) for quick project setup

Complete Langfuse toolkit for AI agent engineering: instrumentation, trace analysis, prompt management, dataset curation, experiments, evaluations, schema validation, and expert optimization advice

Build Langdock integration actions and use Langdock APIs with live documentation fetching. Includes CLI tools for Agent, Knowledge Folder, and Usage Export APIs.

A comprehensive writing assistant with quality loop workflow: deep discovery, voice profiles, iterative self-critique, and publication-ready output

Personal management plugin for daily planning with Linear, German business communication, and YouTrack documentation

Comprehensive toolkit for building MCP servers with the OpenAI Apps SDK. Provides skills, commands, and agents for creating ChatGPT apps with Python and TypeScript.

Structured feedback loop for capturing learnings from plugin usage anywhere and consolidating them into improvements

Long-running agent continuity via Linear + compound loop learning extraction

Personal tracking and goal management system with Supabase integration

Frontend UX evaluation using User Lifecycle Framework and Playwright browser automation

Iterative optimization loops for AI agents using a three-layer framework: (1) Evaluation Infrastructure - dataset, graders, baseline; (2) Optimization Target - goal, constraints, main knob; (3) Optimization Loop - cloud-ready execution prompts. Based on Anthropic's agent eval best practices.