ai-agent-redteam-benchmark-lab

Adversarial containment and security evaluation suite for AI agents.

Download v1.0.0: $199
Future updates: $49 each

You're only charged when you choose to update.

Turn on JavaScript to pay from this page, or pay from the terminal — see Docs.

Buy with your agent

Buy genesiscode2026/ai-agent-redteam-benchmark-lab from X402 Git for me. Listing: https://x402git.com/genesiscode2026/ai-agent-redteam-benchmark-lab How to pay: https://x402git.com/.well-known/agent-skills/buy-a-repo/SKILL.md x402 endpoint: https://x402git.com/api/r/genesiscode2026/ai-agent-redteam-benchmark-lab

What you get

A .zip of the repository at v1.0.0 (12.9 kB)

Size: 16.7 kB
Files: 22
Folders: 2
Dependencies: None
Licence: licence undetermined
Class: Toolkit
Updated: 2026-09-16

Security scan

Secrets: none found
Size limits: within caps
Malware: no dependency manifests to check
Vulnerabilities: no dependency manifests to check
Dependencies: nothing on a direct dependency · 1 in transitive packages
Injection: none found
Capabilities: nothing to declare

2026-09-16: scanned · scanner 2026.09.1

Show report

Languages

JavaScript 100%

FoldersFiles

Top level of the repository
Size	Folder	Inside
4.2 kB	src/	1 source file
1.7 kB	tests/	1 test file
10.8 kB	(root)	15 docs · 3 config files

22 files · 16.7 kB · computed 2026-09-16
Size	Path
59 B	.gitignore
292 B	AGENT_EVAL.md
674 B	ARCHITECTURE.md
221 B	BENCHMARK_REPORT.md
157 B	CHANGELOG.md
276 B	DEPENDENCIES.md
1.1 kB	LICENSE.md
300 B	LIMITATIONS.md
1.7 kB	PREVIEW.md
375 B	QUICKSTART.md
976 B	README.md
291 B	SECURITY.md
225 B	SUPPORT.md
210 B	TEST_REPORT.md
185 B	THIRD_PARTY_NOTICES.md
477 B	THREAT_MODEL.md
195 B	VERIFICATION.md
494 B	commercial-manifest.json
532 B	package.json
2 kB	release-manifest.json
4.2 kB	src/index.mjs
1.7 kB	tests/redteam.test.mjs

computed 2026-09-16 21:28 UTC · analyzer 0.1.0

Already bought this?

View your purchases

README

AI Agent Red-Team Benchmark Lab

Reproducible adversarial red-team evaluation suite and containment benchmark for autonomous coding agents.

Overview

AI Agent Red-Team Benchmark Lab is a verified, clean-room software product built for senior software engineers, quantitative teams, and AI developers. It provides audited contracts, deterministic runtime semantics, and reproducible verification suites.

Core Capabilities

Verified Correctness: Complete test harness with zero external infrastructure dependencies.
Fail-Closed Security: Strict error handling and edge-case boundary enforcement.
Commercial Ready: Fully documented with architecture diagrams, threat models, and quickstarts.

Quickstart

Review QUICKSTART.md for zero-friction setup in under 5 minutes.

Verification

Run tests locally using the standard test command:

npm test

Licensing

Commercial license granted upon x402 purchase. See LICENSE.md.

Security scan report

Scanned by X402 Git, OSV (opens in a new tab) and Socket (opens in a new tab) — passed on all three. Findings from Socket are listed below.

secrets: none found; X402 Git
size limits: within caps; X402 Git
known malware: no dependency manifests to check; OSV
known vulnerabilities: no dependency manifests to check; OSV
dependency behaviour: nothing on a direct dependency · 1 in transitive packages; Socket
instruction smuggling: none found; X402 Git
what it can do: nothing to declare; X402 Git

What this can do

No tools, no hosts, no environment variables, nothing runs at install.

Findings (1)

LOWSocket SBOM Resolver — Socket SBOM Resolver (transitive) — no lockfile, so installs are not reproducible

Releases 1

v1.0.02026-09-16
Initial commercial release v1.0.0