Philipp Schmid's Blog Archive on AI Agents and Gemini (April–September 2026)

LinkPhilschmidResource

The post archive from Philipp Schmid's technical blog outlines articles published between April and September 2026. The primary focus centres on AI agent architectures, the Gemini API ecosystem, and practical tooling integrations.

Key points

  1. Coverage of autonomous agents, subagent design patterns, harness engineering, and recursive self-improvement.
  2. Tutorials on controlling Android devices using Gemini 3.5 Flash and Gemini 3.7 Flash via Computer Use.
  3. Integration guides for Gemini Managed Agents, the Interactions API, and combining tools such as Google Maps and Google Search.
  4. Best practices for using Model Context Protocol (MCP) servers and protecting GitHub tokens within managed agents.
  5. Quick setup tutorials, including deploying OpenClaw 2.0 with Gemini, and methodologies for evaluating agents beyond single prompts.

Why it matters

The archive outlines practical software engineering patterns for modern AI agents in 2026, highlighting real-world automation, OS-level control, and the expanding Gemini developer platform.

Analysis

The catalogued posts demonstrate an industry shift away from basic prompting towards systems engineering for agents. Key recurring themes include multi-agent orchestration, standardized tool connectivity via MCP, and direct operating system execution across Bash and mobile environments.

Mentioned links

Found in

Summary generated by AI (Gemini 3.8 Flash, 27 September 2026). It may contain errors – the original source is authoritative.

Related entries

Similar in content, found by the AI search.

  • Video

    Video:AI Engineer

    Don't Ship Skills Without Evals: Best Practices from Google DeepMind

    Google DeepMind's Philipp Schmid explains why AI agent skills require systematic evaluation suites and how to build lightweight eval harnesses. Drawing on data from SkillsBench and internal practices at Google, he outlines how to write, structure, test, and eventually deprecate skills as foundation models improve.

    AI & LLMs· Talk

  • Link:mobilenext.ai

    Mobile Next: Programmable Mobile Runtime for AI Agents

    Mobile Next provides an infrastructure that allows AI agents to control real iOS and Android devices, simulators, and emulators via API and the Model Context Protocol (MCP). The stack comprises an MCP server, a CLI utility, the Playwright-inspired Mobilewright testing framework, and an on-demand cloud device farm.

    AI & LLMs· Tool

  • Link:Google DeepMind

    Google DeepMind Announces Gemini 3.8 Flash for Agentic Workflows

    Google DeepMind has introduced Gemini 3.8 Flash, an agentic workhorse model designed for complex multi-step reasoning, coding, and knowledge tasks at low latency. The multimodal model supports a 1M token input context alongside extensive tool integration.

    AI & LLMs· Announcement

  • X post:Niels Schmidt

    Mobile Harness: Open-Source iOS and Android Control for AI Agents

    Niels Schmidt has introduced Mobile Harness, an open-source harness enabling coding agents such as Codex or Claude Code to drive real iOS and Android devices locally and in the cloud through a unified API.

    1.3kbookmarks124.6kviews

    AI & LLMs· Announcement

  • Repository:ShawnPana/phone-harness

    Phone Harness: AI Agents Controlling Real Smartphones Without Companion Apps

    Phone Harness is an open-source tool that allows AI agents such as Claude Code or Codex to control physical smartphones. It bypasses jailbreaks, Xcode, and WebDriverAgent by relying on macOS Sequoia iPhone Mirroring for iOS and ADB for Android.

    2.3kstarsPython

    AI & LLMs· Tool

Back to the knowledge base

Let's talk about your project

Get in touch

Locations

  • Mattersburg
    Johann Nepomuk Bergerstraße 7/2/14
    7210 Mattersburg, Austria
  • Vienna
    Ungargasse 64-66/3/404
    1030 Wien, Austria

Parts of this content were created with the assistance of AI.