AI SECURITY INFRASTRUCTUREVERSION 0.1.0 • LOCAL SIMULATION

AI Agent Firewall

Security infrastructure for AI agents, protecting requests before they reach tools, systems and sensitive actions.

Documentation

AI Agent Firewall

A security layer that evaluates AI agent requests before they reach connected tools and systems.

01 — INSPECT

Evaluate incoming AI agent requests.

02 — ENFORCE

Apply security policies and risk-based decisions.

03 — DECIDE

Allow, review, or block requests.

AI AGENT REQUEST
↓
SECURITY ANALYSIS
↓
ALLOWED
REVIEW
BLOCKED
DEMO DATA

SECURITY OPERATIONS WORKSPACE

SECURITY OVERVIEW

TOTAL
--
ALLOWED
--
REVIEW
--
BLOCKED
--
ERROR
--

SYSTEM STATUS

Overall: --
Firewall: --
Policy: --
Audit: --

CASES SUMMARY

Open Cases: --
Closed Cases: --

RECENT ACTIVITY

SECURITY CONTROL CENTER

SYSTEM STATUS: CHECKING...
FIREWALL ENGINE --
POLICY SYSTEM --
AUDIT SYSTEM --
AUDIT CHAIN --
DECISION TRACE --
EXECUTION BOUNDARY --
REQUEST IDENTITY --
REPLAY PROTECTION --
INTEGRATION BOUNDARY --
READ-ONLY

SECURITY ANALYTICS

FIREWALL STATUS: CHECKING...

TOTAL CHECKS

--

ALLOWED

--

REVIEW REQUIRED

--

BLOCKED

--

ERRORS

--

Decision Distribution

Allowed (0%) Review (0%) Blocked (0%) Error (0%)

ALLOWED

--

REVIEW

--

BLOCKED

--

Try your first security check

Enter an AI-agent request. See how ZEQCY evaluates it. Nothing is actually executed in simulation mode.

SECURITY POLICY

POLICY VERSION

--

FINGERPRINT

--

CONSISTENCY

--

DRIFT STATUS

--

POLICY STATE

--
READ-ONLY

SECURITY INVESTIGATION CASES

Case ID Title Status Events Created Actions

SECURITY EVENT EXPLORER

TIME REQUEST ID AGENT TOOL DECISION STATUS DETAILS
NO EVENTS FOUND
Showing 0-0 of 0

How ZEQCY Works

Conceptual integration boundary.

REQUEST

Intercept agent prompt

VALIDATE

Check formatting

ANALYZE

Evaluate intent

POLICY

Match security rules

DECISION

Determine action

REVIEW

Human oversight

Security by Design

✓ Simulation-first evaluation boundary
✓ Original firewall remains authoritative
✓ Deterministic release checksums (SHA-256)
✓ Offline operation — no cloud dependency
Transparency Notice: ZEQCY Version 0.1.0 is a security infrastructure prototype. It focuses on safely evaluating AI agent requests, maintaining immutable audit trails, and preventing execution of unauthorized intents through a protected simulation boundary.

Developer Integration

Integrate the policy engine into your agent framework.

const decision = await FirewallClient.evaluateRequest({
  tool: "DELETE_FILE",
  action: "delete",
  target: "/example.txt"
});

console.log(decision);
{
  "tool": "DELETE_FILE",
  "action": "delete",
  "target": "/example.txt"
}
decision = firewall.evaluate_request({
    "tool": "DELETE_FILE",
    "action": "delete",
    "target": "/example.txt"
})

print(decision)

Security Response

Structured response returned by the firewall security layer.

SIMULATION MODE
status"success"
request_id"req_demo_001"
session_id"session_demo_001"
agent_id"agent_demo"
tool"DELETE_FILE"
action"delete"
target"example.txt"
decision"REVIEW_REQUIRED"
result"Human approval required"
simulationtrue
details"Request requires security review."

Future integration boundary: The frontend expects a structured security response matching the 11-field firewall contract. A backend integration can later provide this response without changing the UI contract.

Documentation & Quick Start

Learn how to use the ZEQCY AI Agent Firewall.

1 Quick Start Guide

STEP 1: Download EXE.
The EXE is the easiest option for most Windows users.
STEP 2: Open the downloaded file.
Run the executable directly.
STEP 3: If using ZIP, extract it first.
Alternatively, extract the ZIP to a local folder and open it.
STEP 4: Launch ZEQCY.
A terminal window will appear to host the local API engine.
STEP 5: Open the security console.
The UI will be accessible locally to provide a safe interaction surface.
STEP 6: Enter a test request.
Try "Open YouTube" as a safe example.
STEP 7: Review the firewall decision.
Understand why the decision was made based on the simulated outcome.

2 Architecture & Simulation

The current release operates strictly in Simulation Mode. The system evaluates AI agent prompts, classifies the risk, and returns a security decision without executing any real-world actions.

  • No Real Execution: The system does NOT run shell commands, delete real files, or access your network.
  • Security Evaluation: It focuses purely on analyzing intent, matching security rules, and providing the correct decision boundary.
  • Authoritative Backend: All decisions are made by the compiled local Python firewall engine, not the web frontend.

3 Understanding Decisions

ALLOWED

The original firewall policy permits the request. In a live system, this would proceed to execution.

REVIEW_REQUIRED

The request requires human oversight and is not executed. It is held for manual approval.

BLOCKED

The firewall identified a high-risk policy violation. The request is permanently blocked.

ERROR

The request could not be safely processed (e.g., malformed data) and is rejected.

Frequently Asked Questions

ZEQCY is a security infrastructure layer that sits between your AI agent and your system. It evaluates the agent's intent before allowing it to execute tools, preventing destructive or unauthorized actions.

No. Version 0.1.0 operates strictly in Simulation Mode. It evaluates intent as text and determines the security decision, but never executes real filesystem or shell operations.

No. The current release is fully local and self-contained. It does not require an internet connection, API keys, or cloud infrastructure to operate.

No. The Windows EXE package bundles the necessary runtime. You can run it directly without installing Python or external dependencies.

All simulated security evaluations are logged to an immutable audit trail. You can view them in the local Security Console by clicking the Events or Cases tabs.

Currently, only Windows x64 is supported. macOS and Linux support is planned for future releases.

Download AI Agent Firewall

Secure your AI agent workflows with the ZEQCY AI Agent Firewall.

Current Release

ZEQCY AI Agent Firewall — Version 0.1.0

Build: Simulation / Prototype  •  Execution: Local only  •  External API: Not connected

SHA-256 (EXE): DA0B7265F0DA44C57BEC4D346502601D5990DA782B3C629B7778D845D19D6D6D
SHA-256 (ZIP): CD9DA3EA8F4C80E401F3B067928C9C3DAAB6CC3BADFB584AE09592F4D27EE7A3

Platform

Windows x64 macOS — Coming Soon Linux — Coming Soon

Only the Windows package is currently available. macOS and Linux are not yet supported.

System Requirements

  • OS: Windows 10 or later (64-bit)
  • Browser: Any modern browser (Chrome, Firefox, Edge) for the Security Console
  • Network: No internet required — fully local operation
  • Disk: ~50 MB free space
  • Python: Not required — runtime is bundled in the EXE

Installation & Launch

  1. Download the EXE or ZIP package above.
  2. Extract (ZIP only): Right-click → Extract All → choose a folder.
  3. Run ZEQCY-AI-Agent-Firewall.exe — no installation wizard needed.
  4. Open your browser and navigate to http://127.0.0.1:8080 for the Security Console.
  5. Submit a test request in the Security Console simulator.
  6. View the security decision — ALLOWED, REVIEW REQUIRED, or BLOCKED — returned by the firewall.

The firewall API runs locally at 127.0.0.1:8081. No internet connection is required.

What's Included

  • ZEQCY AI Agent Firewall engine (bundled runtime)
  • Security Console website (local HTML/JS/CSS)
  • Security Dashboard, Event Explorer, Investigation View, Case Management
  • Policy Visibility & Control Center
  • Simulation-mode execution boundary (no real system commands)

Current Limitations

  • Tool execution is simulated — no real filesystem, network, or shell commands are executed
  • No external AI API or cloud dependency
  • No payment or subscription required
  • No public Agent API — local operation only
  • No authentication system in this prototype release
  • Windows x64 only in this release

Security & Integrity

ZEQCY AI Agent Firewall currently operates as a simulation-first security prototype. No real system commands, network actions, email actions, or money transfers are executed. Verify the SHA-256 checksum above before running the downloaded file.

Product AI Agent Firewall
Company ZEQCY
Version 0.1.0
Build Simulation / Prototype
Execution Local
External API Not connected
View Documentation Explore Developer Portal