Philipp Schmid

mcp-cli is a lightweight CLI tool that enables dynamic discovery of Model Context Protocol servers, reducing token consumption by up to 99% while making tool interactions more efficient for AI coding agents.

“Author of "MCP CLI" in Philipp Schmid”

“Author of "Context Engineering for AI Agents"”

Deep Agents (Agents 2.0) that combine explicit planning, hierarchical delegation, persistent memory, and adaptive execution can handle multi-step complex tasks spanning days, whereas shallow loop-based agents fail on tasks requiring more than 50 steps.

“Author of "Build agents that handle days-long tasks without collapsing" in Philipp Schmid Blog”

“Author of "Subagents Cut Context Pollution In Multi-Tool Agents"”

“Created a skill for the Gemini API CLI that gives agents command-line access to call Gemini models for images, TTS, video and files.”

“Shared update on MCP roadmap with upcoming support for long-running workloads, HTTP protocol, and progressive discovery.”

“Co-authored announcement about Gemini API Managed Agents updates including Gemini 3.6 Flash and environment hooks.”

“Showed how agent skills in markdown allow agents without code, reducing orchestration complexity.”

“Schmid announced new capabilities for Managed Agents in Gemini API including background execution and remote MCP integration.”

“Co-authored Google's Gemini Managed Agents guide introducing the Antigravity agent and managed agent infrastructure.”

“Authored article on best practices for building MCP servers and why most MCP servers fail to deliver value for AI agents.”

“Byline author making 8 predictions for 2026 AI including edge-first personal agents and generative UI advancement.”

Google introduces the Interactions API, a unified interface replacing stateless calls with agent-native workflows that supports complex context management, tool use, and built-in agents like Gemini Deep Research.

“Co-author introducing Google's Interactions API for managing long-running agent interactions with models and agents.”