RedHub Prompt Injection Red Team Kit

Spec sheet

Purpose

Runs 15 OWASP-mapped probes against an LLM app, scores its resilience, and returns a ship, hold, or fix verdict.Read 7 Oct 2026 · redhub.ai

Intended users

The maker describes it for founders and engineers shipping chatbots, copilots, RAG apps, and agents, and for consultants adding security checks to AI projects.Read 7 Oct 2026 · redhub.ai

Support

RedHub’s Help Center says customers can email [email protected] for help with products and downloads.Read 7 Oct 2026 · redhub.ai

Headquarters

Ormond Beach, Florida, USARead 28 Sep 2026 · redhub.ai

Maker

RedHub AI LLC was founded by Todd Brooks; the company identifies Ormond Beach, Florida, USA as its location.Read 7 Oct 2026 · redhub.ai

Authorized use

The maker says to run the defensive kit only against systems you own or are explicitly authorized to test.Read 7 Oct 2026 · redhub.ai

CI gate

The kit provides tunable severity-weighted thresholds, ship/hold/fix verdicts, exit codes 0/1/2, and a drop-in GitHub Actions workflow.Read 7 Oct 2026 · redhub.ai

Detection

The detector uses benign markers and a planted canary with exact-match detection; an optional bring-your-own LLM judge is supported but not required.Read 7 Oct 2026 · redhub.ai

Integration

To connect an app, implement an adapter function that accepts a prompt and optional context and returns the app’s reply; generic, callable, and HTTP adapter templates are included.Read 7 Oct 2026 · redhub.ai

Languages and dependencies

The kit provides Python with zero dependencies and dependency-free TypeScript ESM with full parity; both run offline and produce identical verdicts.Read 7 Oct 2026 · redhub.ai

Limits

A SHIP verdict means resilience to the tested patterns, not immunity; the maker says the kit is not a runtime firewall.Read 7 Oct 2026 · redhub.ai

Privacy of test runs

The maker says the kit runs entirely offline and sends nothing anywhere.Read 7 Oct 2026 · redhub.ai

Probe coverage

The probes cover OWASP LLM01, LLM02, LLM05, LLM06, and LLM07, including direct and indirect injection, disclosure, output handling, excessive agency, and system-prompt leakage.Read 7 Oct 2026 · redhub.ai

Reports and guidance

The kit includes an HTML report, OWASP-mapped remediation guide, comparison document, security-and-ethics charter, and three mock targets.Read 7 Oct 2026 · redhub.ai

Plans and prices

Every RedHub Prompt Injection Red Team Kit plan and price →

Where Specifiction places it

• AI Red Teaming Tools#21 of 277.1