mcp-graph
Open-source CLI (v7.0) that converts PRDs into persistent execution graphs, coordinates 5 MCPs, reduces tokens by up to 85%, and includes 155 multimodal skills. Anti-vibe-coding with Harness Engineering.
// Hi, I'm |
Full-Stack & AI Engineer with 15+ years of experience. Specialized in GenAI, LLMs, RAG, DevOps, MLOps and Cloud-Native architectures. Currently Software Architect at Vivo (Telefonica Brazil).
Open-source CLI (v7.0) that converts PRDs into persistent execution graphs, coordinates 5 MCPs, reduces tokens by up to 85%, and includes 155 multimodal skills. Anti-vibe-coding with Harness Engineering.
Free, open-source temporary email platform for developers. Disposable emails, REST API, webhooks and automatic encryption.
It is not a bubble. AI's individual gain is real and measured, but it leaks on the way to the company's bottom line. With data from WRITER, MIT, Gartner, Goldman, Forrester and Google DORA, I show where the money disappears and what separates the 29% who capture value.
A real reward hacking case in mcp-graph: how feature-depth became an optimization target and why static analysis needs behavioral validation in agentic loops.
AISE is the emerging discipline of 2026. I prove with the repo and the DORA Report 2025 that mcp-graph v12 is the runtime layer where Specification-Driven Development meets Context-Driven Engineering.