> Source: https://botzr.com/
> Part of Botzr Research — https://botzr.com
> Last updated: 2026-08-11

Offensive Security · AI Agents Est. 2023 · Oklahoma US

# botzr research.

Independent offensive security research: AI agent systems, MCP infrastructure, and the memory they trust. The work happens in public. Findings are the product.

00 / RED TEAM

## [ECHOS](https://botzr.com/programs/echos)

Adversarial security research against self-hosted MCP servers and the agent infrastructure around them. Real vulnerabilities, disclosed responsibly. No targets named before disclosure.

[About ECHOS→](https://botzr.com/programs/echos)

First advisories Q4 2026

01 / TOOLING

## [AEGIS](https://botzr.com/programs/aegis)

A Rust and WASM sandbox for detonating and constraining untrusted MCP tool calls. The harness the red team work runs inside.

[About AEGIS→](https://botzr.com/programs/aegis)

In development

02 / BENCHMARK

## [dreamd](https://dreamd.dev)

A reproducible benchmark for how AI memory systems handle facts that change, including facts an attacker changes. dreamd is one row in the table, not the headline.

[dreamd.dev→](https://dreamd.dev)

Public August 2026

03 / PRIMER

## [Memory as attack surface](https://botzr.com/research/state-drift)

What happens to an AI system's stored beliefs when the facts underneath them change, or are made stale on purpose: what the field has measured so far, and what remains untested.

[Read the primer→](https://botzr.com/research/state-drift)

Background

04 / PRINCIPAL

## Uveddi

Austin Green. Independent researcher and systems engineer in Oklahoma. 25 years of systems judgment, applied in public. The work speaks; the person is secondary.

[uveddi.org](https://uveddi.org) [@wgi\_dev](https://x.com/wgi_dev)

Oklahoma, US