Projects

The tools and applications I’ve been working on, with source code, examples, and some context.

Ongoing

Checking whether an AI coding agent has enough evidence to call its work done.

Changing one historical input, then replaying what happens to the transactions that follow.

Spending permissions with limits, revocation, and a receipt you can inspect.

An AI portfolio experiment with recorded proposals, validation, and execution evidence.

Telegram bots and agent automation, where the hands-on work began.