Adversarial Testing Layer

Moriarty

Pressure-test the workflow before the workflow embarrasses you.

Moriarty is the red-team and failure-seeking layer in the Sherlock product set, built for prompt injection, policy evasion, tool misuse, and remediation verification work.

Discuss this product View the stack
Scenario packsRules-of-engagement checklistsWorked support-assistant red-team example

Status

Open-source red-team starter kit

Domain

Adversarial testing and remediation verification

Primary outcome

A clearer view of exploit paths, severity, and whether fixes actually worked

Best paired with

Bid support, rapid build hardening, or pilot rescue assurance

Overview Artifacts Workflow Product intake

Adversarial pressure where weak assumptions usually hide

Moriarty helps teams test the edge cases they usually postpone: prompt injection, identity confusion, unsafe tool use, and policy bypass attempts. It is designed to turn red-team work into a structured delivery artifact instead of a vague security note.

Who this product is built for

The outputs teams actually use

These are the delivery artifacts the repo is designed to produce, not just the internal implementation detail.

Artifact

Assessment plan

Defines scope, rules of engagement, coverage, and operational boundaries before testing starts.

Artifact

Attack report

Documents the exploit path, severity, impact, and recommended mitigation.

Artifact

Remediation verification

Proves whether the shipped fix actually closes the failure path.

How it gets used in real delivery

Each product is designed to slot into a fixed-scope Baker Street engagement rather than sit as a disconnected side project.

Step 1

Set scope, boundaries, and the attack classes that matter most.

Step 2

Run scenario packs against the workflow and record exploit evidence.

Step 3

Verify mitigations after fixes land so the issue is not merely reworded away.

Use it alongside the rest of the Baker Street system

Need to know how the workflow fails before a client does?

Moriarty is built for pre-production hardening, pilot assurance, and any team that needs a structured adversarial pass instead of a vague security conversation.