# Claude Agents SDK Projects at AI Tinkerers

> Canonical HTML: https://aitinkerers.org/technologies/claude-agents-sdk
> Markdown URL: https://aitinkerers.org/technologies/claude-agents-sdk.md
> Technology record last updated: 2026-03-02T18:20:43Z
> Generated: 2026-09-20T20:45:45Z

The Claude Agents SDK is a Python and TypeScript framework that grants Claude agents terminal access, file system operations, and tool orchestration for autonomous, complex digital workflows.

This SDK is your direct path to building powerful, general-purpose AI agents; it moves beyond text-only responses. It provides a robust framework that manages the agentic loop (Gather Context, Take Action, Verify Work), significantly reducing boilerplate code for multi-step reasoning. Developers use the Python or TypeScript SDK to equip agents with built-in tools like Read, Write, and Bash, enabling them to autonomously perform tasks such as code review, data analysis, and enterprise automation. The system supports advanced features like subagents for parallelization and integrates custom tools via the open Model Context Protocol (MCP).

- Official technology site: https://docs.claude.com/docs/agent-sdk-overview
- Public AI Tinkerers demos and talks: 4
- Result page: 1 of 1

## Recent Public Talks and Demos

### [Building a UI that builds itself](https://copenhagen.aitinkerers.org/talks/rsvp_CD2l5EihkZY)

it's a UI layer for agent harnesses where agents (Claude/Codex/Hermes/OpenClaw) can work and build personal software by extending their own UI. I'll cover why we decided work on this idea, some existing approaches, how we organized the CLI/Skill and what design guidelines we have written.

- Event context: September Demo Night — 2026-09-16 — Copenhagen
- Public talk page: https://copenhagen.aitinkerers.org/talks/rsvp_CD2l5EihkZY

### [Reverse Ralph Loops](https://columbus.aitinkerers.org/talks/rsvp_GkXWwT0FF3A)

I built an app that runs a reverse Ralph Wiggum loop over an existing codebases to create a clean room specification of features or whole applications. You can then use these specifications to recreate the software for yourself. I wanted a deep research tool for my coding agents. So I looked at the leaderboard for deep research agents and picked the top open source project. I then ran an LLM in a loop to work on one aspect of the specification on each iteration until the full specification was created for the deep research functionality. I then used the specification with a loop that handled one implementation task from the spec on each iteration to create my own deep research CLI.

- Event context: AI Tinkerers - Columbus April Meetup — 2026-04-06 — Columbus
- Public talk page: https://columbus.aitinkerers.org/talks/rsvp_GkXWwT0FF3A

### [From ServiceNow Ticket to Terraform Pull Request using Skills, MCP with Coding Agents](https://nashville.aitinkerers.org/talks/rsvp_3hn_1rgR6k0)

Most coding agents are really good at generating infrastructure code from scratch. But how to steer them to follow existing Terraform conventions in the form of base modules: encoding security policies, tagging standards for cost allocation, naming conventions, and compliance guardrails into reusable building blocks. These modules are how we enforce consistency across environments, track cloud spend back to business units, and pass regulatory audits. In this talk, I'll show you how to use skills to define a better infrastructure provisioning workflow. One that steers coding agents to first discover your existing conventions, understand the risks, implement changes within those guardrails, and provide full visibility into every decision made along the way. This gives cloud operations teams the ability to orchestrate infrastructure requests with human-in-the-loop validation at every stage, see huge productivity gains The system uses coding agents designed to work directly on a filesystem—reading and modifying code, running bash commands and interacting with real repositories. Built using Agent Skills and the Claude Agents SDK, this talk walks through how the workflow is structured, what works today. I'll also cover how this approach is agnostic to agent runtimes and model providers - built on open standards (MCP + Skills), so you're not locked into any single vendor and how the same pattern extends to other tasks across the enterprise beyond infrastructure provisioning.

- Event context: AI Tinkerers Nashville – Software Development in the Age of AI — 2026-03-25 — Nashville
- Public talk page: https://nashville.aitinkerers.org/talks/rsvp_3hn_1rgR6k0

### [Metadata Driven AI Pipelines](https://seattle.aitinkerers.org/talks/rsvp_3eYiO5c3fGw)

Long running tasks are always a balance of control and agency. As we slowly move away from the CLI to task managers (Jira, Linear, Beads, etc) we will need to figure out ways to find a good balance between control and agency. If control is pure code and agency is purely agentic, then AI pipelines look to be something in between. Code that executes and validates LLM output, but relies on the intelligence of the model to do a majority of the work. I will show how I am designing the AI pipelines and using metadata from a task manager (Linear in this case) to control which pipeline executes for different types of tasks (bugs, features, etc)

- Event context: AI Tinkerers Seattle Meetup: Dev Tools Track — March 9th, 2026 — 2026-03-10 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_3eYiO5c3fGw

## Related Technologies

- [FastAPI](https://aitinkerers.org/technologies/fastapi) ([Markdown](https://aitinkerers.org/technologies/fastapi.md)) — 181 public demos
- [Python](https://aitinkerers.org/technologies/python) ([Markdown](https://aitinkerers.org/technologies/python.md)) — 662 public demos
- [Agent Skills](https://aitinkerers.org/technologies/agent-skills) ([Markdown](https://aitinkerers.org/technologies/agent-skills.md)) — 4 public demos
- [Anthropic API](https://aitinkerers.org/technologies/anthropic-api) ([Markdown](https://aitinkerers.org/technologies/anthropic-api.md)) — 66 public demos
- [Beads](https://aitinkerers.org/technologies/beads) ([Markdown](https://aitinkerers.org/technologies/beads.md)) — 7 public demos
- [Claude Opus](https://aitinkerers.org/technologies/claude-opus) ([Markdown](https://aitinkerers.org/technologies/claude-opus.md)) — 18 public demos
- [Hermes](https://aitinkerers.org/technologies/hermes) ([Markdown](https://aitinkerers.org/technologies/hermes.md)) — 4 public demos
- [Jinja](https://aitinkerers.org/technologies/jinja) ([Markdown](https://aitinkerers.org/technologies/jinja.md)) — 1 public demo
- [Jira](https://aitinkerers.org/technologies/jira) ([Markdown](https://aitinkerers.org/technologies/jira.md)) — 4 public demos
- [Linear](https://aitinkerers.org/technologies/linear) ([Markdown](https://aitinkerers.org/technologies/linear.md)) — 9 public demos
- [OpenClaw](https://aitinkerers.org/technologies/openclaw) ([Markdown](https://aitinkerers.org/technologies/openclaw.md)) — 50 public demos
- [Ralph Wiggum loops](https://aitinkerers.org/technologies/ralph-wiggum-loops) ([Markdown](https://aitinkerers.org/technologies/ralph-wiggum-loops.md)) — 1 public demo
- [React](https://aitinkerers.org/technologies/react) ([Markdown](https://aitinkerers.org/technologies/react.md)) — 219 public demos
- [Vercel fx](https://aitinkerers.org/technologies/vercel-fx) ([Markdown](https://aitinkerers.org/technologies/vercel-fx.md)) — 1 public demo