Live Demos

See OpenA2A in Action

Terminal recordings, video walkthroughs, and quick-start commands for every project in the OpenA2A security platform.

Secretless AIHackMyAgentAIMARPOASBDVAA

Secretless AI

Credential management for AI coding tools

Keeps API keys, tokens, and credentials out of AI context windows. Auto-detects Claude Code, Cursor, Windsurf, Copilot, and more. One command to protect everything.

Auto-detects 6 AI coding tools
49 credential patterns scanned
MCP server secret encryption
Shell profile diagnostics
Secretless AI terminal demo
Quick Start
$ npx secretless-ai init
Learn MoreGitHubnpm

HackMyAgent

Security scanner

Scans and hardens AI agent setups with 147 security checks across 30 categories. Includes adversarial red-teaming with 75+ attack payloads and one-click auto-fix.

147 security checks, 30 categories
75+ adversarial attack payloads
Auto-fix with backup and rollback
SARIF, HTML, JSON output formats
HackMyAgent terminal demo
Quick Start
$ npx hackmyagent secure .
Learn MoreGitHubnpm

AIM

Identity and access management for AI agents

Ed25519 cryptographic identity, real-time trust scoring, and capability-based access control. Register an agent with one line of Python. Dashboard shows everything in real time.

One-line agent registration
Ed25519 cryptographic identity
Real-time trust scoring dashboard
JIT access approval for critical actions
Quick Start
$ pip install aim-sdk
Learn MoreGitHub

ARP

Agent Runtime Protection

3-layer intelligence stack monitors AI agent behavior at runtime. Detects prompt injection, data exfiltration, and MCP exploitation. Budget-controlled LLM-assisted analysis.

3-layer intelligence (rules, stats, LLM)
20 built-in threat detection patterns
HTTP reverse proxy with AI-layer scanning
Budget-controlled at $5/month default
terminal
$npx arp-guard startMonitor
$npx arp-guard tail 10Events
$npx arp-guard budgetBudget
Quick Start
$ npm install @opena2a/arp
Learn MoreGitHubnpm

OASB

Open Agent Security Benchmark

Standardized benchmark for evaluating agent security products. 222 tests covering process, network, filesystem, AI-layer, and intelligence detection. Maps to MITRE ATLAS and OWASP.

222 standardized attack scenarios
10 MITRE ATLAS techniques covered
Atomic, integration, baseline, E2E tests
Full suite runs in under 5 seconds
terminal
$npm testRun 222 tests
$npm run test:integrationAttack chains
$npm run reportReport
Quick Start
$ npm install @opena2a/oasb
Learn MoreGitHubnpm

DVAA

Deliberately vulnerable AI agents for security training

10 intentionally vulnerable AI agents across API, MCP, and A2A protocols. CTF challenges, live attack log, prompt playground. The DVWA of AI agents.

10 vulnerable agents (4 security levels)
API, MCP, and A2A protocol coverage
CTF challenges with 2,550 total points
Built-in prompt playground
terminal
$docker pull opena2a/dvaaPull
$docker run -p 3000-3006:3000-3006 -p 9000:9000 opena2a/dvaaRun
$open http://localhost:9000Dashboard
Quick Start
$ docker pull opena2a/dvaa
Learn MoreGitHub

Full Video Walkthrough

Watch the complete AIM platform demo including agent registration, MCP server attestation, trust scoring, and SDK integration.

Watch on YouTubeView on GitHub