Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
Skills: Proprietary Library Coding
Learn how to make coding agents use proprietary libraries with "Skills." This talk covers structuring and creating these lazy-loaded instructions, with a live demo of building an AI chat app.
Making coding agents use proprietary/internal libraries correctly has been an ongoing frustration of mine. Zero examples in the LLM’s training set, no online docs, and internal docs are sparse and lag behind. Not a great environment for a coding coding. Modern “ripgrep” agents work occasionally (when they trace into the source), but are still pretty hit’n’miss.
Two weeks ago I found a solution: Skills. Skills are, in a nutshell, lazy-loaded instructions for your coding agent. I’ve successfully used them to create guides on how to use proprietary libraries and they have worked pretty well for me so far.
In my talk I will show you how I structure them and how I think through their creation. If the WiFi is stable, I will also show you an example where I will ask a coding agent to one-shot a small AI chat app using a custom agent SDK that I am building.
- PythonPython: The high-level, general-purpose language built for readability, powering everything from web backends to advanced machine learning models.Python is the high-level, general-purpose language prioritizing clear, readable syntax (via significant indentation), ensuring rapid development for any team . Its ecosystem is massive: use it for robust web development with frameworks like Django and Flask, or leverage its power in data science with libraries such as Pandas and NumPy . The Python Package Index (PyPI) provides thousands of community-contributed modules, offering immediate solutions for tasks from network programming to GUI creation . The language is actively maintained by the Python Software Foundation (PSF), with the stable release currently at Python 3.14.0 (as of November 2025) .
- CursorThe AI-native code editor designed for high-velocity development through deep LLM integration.Cursor is a fork of VS Code that embeds AI directly into the development workflow while maintaining full extension compatibility. It leverages models like Claude 3.5 Sonnet and GPT-4o to power features such as Cmd+K for inline edits and Cmd+L for codebase-wide chat. By indexing local files, Cursor provides precise context for its predictive 'Tab' completions and multi-file 'Composer' mode. This setup allows engineers to move from high-level intent to functional code without leaving the editor or losing context.
- MarkdownMarkdown is a lightweight markup language that converts plain text into structured HTML using simple punctuation.John Gruber released Markdown in 2004 to simplify web writing for non-coders and developers alike. It uses intuitive syntax (like # for headers or * for lists) to produce clean XHTML. This standard now powers documentation on GitHub, discussions on Reddit, and static sites via Jekyll. By prioritizing readability, Markdown ensures that a source document remains legible even before it hits a browser.
- SkillsThe essential platform for building in-demand AI and technical capabilities: access free courses, skill badges, and industry-recognized Google Cloud certifications.Skills (Google) delivers targeted, high-impact technical training to close the global skills gap. The platform offers extensive no-cost options for all learning levels, including hundreds of courses focused on AI and technical fundamentals. Users earn shareable digital credentials: complete a certificate learning path for a full certification, or finish a series of courses to secure a Skill Badge, proving practical, technical expertise. This is the direct path for individuals and Google Cloud partners to validate knowledge and advance their careers with industry-recognized credentials.
- Agent SDKDeploy autonomous, multi-step AI agents fast: the Agent SDK provides a lightweight, Python-first framework for LLM orchestration, tool integration, and enterprise-grade tracing.The Agent SDK is your production-ready framework for building complex, autonomous AI agents. It simplifies LLM orchestration (supporting OpenAI, Anthropic, and others) by managing the full execution loop, tool invocation, and multi-agent handoffs. Developers use Python functions to create tools (e.g., `WebSearchTool`, `CalendarTool`), which the agent autonomously calls to complete tasks. Key features include built-in tracing for debugging and optimization, plus configurable Guardrails for input/output validation. This is the efficient path to deploying reliable agents in production.
Related projects
Skills are eating AI agents
Orange County
Learn how skills are now central to AI agents, enabling reliable end-to-end automation of business workflows. See concrete…
Is your skill still doing the right thing?
Poland
Learn about a new tool to test if your SKILL.md is followed by popular agents. Verify skill loading,…
Coding with AI: What Works & What Doesn't
Manchester Nh
Learn practical AI coding assistant dos and don'ts, avoiding slowdowns. See a live demo of a self-learning agent…
Methodologies for the Optimization of Developers' Workflow with AI Agents
Pereira
Learn to optimize AI developer workflows using Plugins, MCP servers, and Custom Skills. Move beyond vibe coding to…
Agentic learning via skills
New York City
See how an agent learns new information from external data, solves problems, and updates itself, enabling faster experimentation…
Code > Prompts: Why Your Agent Should Think in Code, Not English
Seattle
See how agents can emit typed Python execution plans instead of English reasoning. This talk demonstrates a framework…