Disconnected testing
Prompts, tools, and framework notes live in separate places with no consistent execution method.
AI SECURITY OPERATING ASSETS
Stop starting every AI red-team assessment from scratch.
A complete implementation kit for planning, executing, documenting, and reporting authorized adversarial assessments—with 20 interactive starter tests, a 150-case downloadable library, and a private local-first workspace.
150DOWNLOADABLE TESTS
20INTERACTIVE TESTS
16WORKSPACE MODULES
One-time Individual license · Twelve months of v1.x updates · Authorized testing only.
CGN // ASSESSMENT CONTROLREADY
AIPIRAGTOOLMEM
ASSESSMENT SIGNAL
CONTROL STATUS4 findings require reviewEvidence package synchronized
600+
GitHub stars and growing
MICROSOFTOPENAIGOOGLENVIDIAAMAZONCISCO
THE IMPLEMENTATION GAP
A folder of jailbreak prompts is not a red-team program.
The free guide gives you the methodology. RedTeamKit gives you the operating assets to apply it consistently.
Weak evidence
Interesting behavior becomes difficult to reproduce, score, remediate, and explain to decision-makers.
No operating standard
Every assessment starts from zero, creating inconsistency across testers, systems, and releases.
FREE GUIDE VS. REDTEAMKIT
The guide explains the discipline. The kit operationalizes it.
Learn the methodology
- Public field and framework reference
- General attack and control guidance
- Implementation recommendations
- Free and continuously updated
Run the assessment
- 150 structured downloadable test cases
- 20 interactive starter tests
- 16 local-first workspace modules
- 12 editable source templates
- Northstar completed assessment
- Versioned product updates
A REPEATABLE SYSTEM
From authorization to retest.
A seven-stage operating model keeps assessments risk-based, reproducible, and useful to technical and executive audiences.
STAGE 01
ENGAGEMENT CONTROL
Define the rules before testing begins.
Establish the target, authorization, boundaries, stakeholders, data handling, prohibited activities, and emergency contacts.
- Rules of Engagement
- Assessment scope
- Testing authorization
WHAT YOU RECEIVE
Professional assets, ready to use.
Every component moves the assessment forward—not another document to read.
01
Structured attack library
Twenty starter tests run interactively in the workspace. The complete 150-case Excel and CSV library covers prompt, RAG, tool, MCP, memory, multi-agent, and resilience risks.
TEST IDCATEGORYRISK
PI-001InjectionHIGH
RAG-014RetrievalHIGH
MCP-011Tool useCRIT
MEM-008MemoryMED
Quickstart playbook
Scope, map, prioritize, test, measure, report, and retest without inventing the workflow.
Assessment workspace
Manage system intake, checklists, test execution, findings, owners, remediation, and reports locally.
12 editable templates
Rules of Engagement, threat model, finding report, executive summary, security card, and more.
Northstar example
A completed fictional agent assessment showing how the assets work together.
12 months of updates
New test cases, framework mapping revisions, templates, and workflow refinements.
SEE THE SYSTEM
Built for the work behind the report.
Explore the core operational views included in the kit.
TESTDOMAINRISKSTATUS
Instruction hierarchy overridePromptHIGHReady
Indirect document injectionRAGHIGHReady
Unauthorized tool transactionMCPCRITReady
FINDINGSEVERITYOWNERSTATUS
Indirect injection via HTMLHIGHM. ChenMitigating
Tool over-permissioned for writesCRITR. OkaforOpen
Cross-tenant retrievalCRITS. PatelRetest
METRICVALUETRENDTARGET
Attack success rate12.8%▼ 4.2<10%
Open criticals2▼ 10
Regression pass96%▲ 3100%
Local-first by design.Assessment data stays in the buyer’s browser unless exported. The workspace must not contain unredacted production secrets, regulated data, or customer records.Open read-only demo →
LICENSED FOR HOW YOU WORK
Choose your level of access.
One-time purchase. Twelve months of product updates included.
Build your capability.
For one practitioner using the kit for personal, portfolio, or individual internal work.
$249
- 150 structured tests
- 20 interactive starter tests
- 16-module workspace
- 12 editable templates
- Northstar example
- 12 months of updates
Use it with clients.
For consultants and service providers delivering authorized paid assessments.
$499
- Everything in Individual
- Commercial-use license
- Client discovery questionnaire
- Proposal and statement of work
- Client-ready reports
- Pricing worksheet
Standardize your practice.
For internal security, AI, product, risk, or assurance teams of up to ten users.
$1,499
- Everything in Individual
- Up to 10 internal users
- Shared workbook assets
- Program charter
- Rollout checklist
- Team governance pack
START WITH A FIELD SAMPLE
Get 10 structured AI red-team tests.
Preview the case format, expected secure behavior, evidence requirements, and control guidance before purchasing the complete library.
Explore the demoBEFORE YOU BUY
Frequently asked questions.
Is this a course or certification?
No. It is a professional implementation kit delivered as downloadable files plus a local-first assessment workspace. It has no video lessons, graded assignments, or certification.
What is interactive versus downloadable?
The workspace includes 20 interactive starter cases. The complete 150-case library is delivered in editable Excel and CSV formats.
Can I use it with clients?
Yes, with the Consultant License, during explicitly authorized engagements. The source files may not be redistributed or resold.
Where is assessment data stored?
Version 1 stores workspace state in the buyer’s browser. It is not a hosted multi-tenant SaaS service. Buyers must create backups and avoid unapproved sensitive data.
Does it authorize testing?
No. Written authorization, exact scope, approved data handling, and applicable law and policy are always required.
What happens after purchase?
Stripe securely processes payment, billing, tax details, and receipts. After payment is confirmed, RedTeamKit emails a six-digit verification code so the buyer can activate a password-protected account and access the licensed dashboard and tier-specific downloads.
FROM GUIDE TO OPERATING CAPABILITY
Your next assessment should begin with a system—not a blank document.
Get the assets needed to conduct structured, authorized, and defensible AI red-team assessments.