A free day of training. A shared mission.

Secure coding
for America.

With AI. With Jim Manico.

A free, full-day workshop on Claude Code, Codex, and secure AI coding for the people who build and protect America’s public services.

Save your free seat Explore the day

Nonpartisan. Open to every political affiliation. Free to attend.

Claude Code + Codex8-hour AI coding workshopMCP securityOne shared mission

01 / Who this is for

You serve the public.
We’re here to help you
secure the code.

Public service depends on software people can trust. This class is for the developers, security engineers, IT staff, and technical leaders who help make that possible.

Federal government

Agency teams, civilian staff, military personnel, and the technologists supporting federal public services.

State & local government

State, city, county, municipal, and tribal teams building and maintaining the services communities rely on.

Political party organizations

Technology and security staff across all U.S. political parties. Every affiliation is welcome.

The public-service community

Government contractors, civic technology groups, and related organizations working to support America.

Designed around government and civic technology work. Open to anyone who wants to learn.

02 / Your full-day workshop

From a security spec
to code you can verify.

8hours / full-day workshop
including a one-hour lunch

Spend the day working with Claude Code, Codex, and AI coding techniques across the secure development lifecycle—from setting up your agents to shipping and maintaining the result. Follow one training application using synthetic data and realistic public-service scenarios. Watch the live demonstrations or code along.

  1. Specify

    Use spec-driven development (SDD) to define roles, security requirements, abuse cases, and testable acceptance criteria.

    Output: a security specification

  2. Threat model

    Map data flows and trust boundaries. Identify threats and turn the highest risks into development tasks and negative tests.

    Output: a practical threat model

  3. Build

    Use Claude Code and Codex to implement small, reviewable changes with authorization checks, secure defaults, and regression tests.

    Output: an implementation with tests

  4. Review

    Review the code, reproduce potential vulnerabilities, fix confirmed issues, and verify the changes against the original requirements.

    Output: a reviewed change with evidence

The goal: a repeatable secure development process you can bring back to your team.

03 / Throughout the workshop

Secure the tools.
Verify the results.

Before you build

Operate coding agents safely

Configure Claude Code and Codex with deliberate permissions, sandboxing, project instructions, and secrets handling. Recognize prompt injection in repository files and tool output.

A full hour on MCP

Connect tools with control

Use Model Context Protocol (MCP) with your coding agents. Connect a training server and explore tool permissions, scoped credentials, untrusted responses, and authorization failures.

45 minutes / Verification

Prove the code is secure

Combine AI review with security tests, static analysis, dependency scanning, and secret detection. Reproduce findings and verify the fixes.

30 minutes / Responsible delivery

Ship and maintain responsibly

Bring AI-assisted development into your team with human review, dependency policies, audit evidence, rollback, and a practical adoption plan.

04 / Your day, planned

A full day of secure AI coding.

Wednesday, November 11, 2026
All times Eastern (EST / UTC−5)
8-hour workshop, including a 1-hour lunch

  1. 9:00–9:45 AM
    AI coding foundations · 45 minutes

    Secure your AI coding environment

    Configure Claude Code and Codex permissions, sandboxing, secrets handling, and project instructions; recognize prompt injection in repository files and tool responses.

  2. 9:45 AM–12:30 PM
    SDD, threat modeling & coding · 2 hours 45 minutes

    Specify, threat model, and build

    Use Claude Code and Codex to take one feature from its security specification through threat modeling and implementation.

  3. 12:30–1:30 PM

    One-hour lunch break

    Step away, recharge, and come back ready to review.

  4. 1:30–2:45 PM
    Testing & secure code review · 1 hour 15 minutes

    Finish, test, and review

    Use Claude Code and Codex to complete the feature, perform secure code review, reproduce findings, and add security regression tests.

  5. 2:45–3:45 PM
    MCP security · 60 minutes

    MCP: connect tools without surrendering control

    Connect your coding agents to a small MCP server, understand its trust boundaries, and exercise its authorization controls.

  6. 3:45–4:30 PM
    AI coding verification · 45 minutes

    Prove the code is secure

    Combine Claude Code and Codex review with security tests, static analysis, dependency scanning, and secret detection.

  7. 4:30–5:00 PM
    Responsible AI delivery · 30 minutes

    Ship and maintain responsibly

    Establish human review of AI-generated changes, dependency policies, audit evidence, rollback, and a practical team adoption plan.

Jim Manico, founder of Manicode Security

05 / Your instructor

Jim Manico

Founder, Manicode Security

Jim teaches developers how to build more secure software. He leads the OWASP Cheat Sheet Series and is the author of Iron-Clad Java: Building Secure Web Applications.

For this free class, he brings that practical secure coding approach to AI-assisted development—with examples designed for the people supporting America’s public services.

Meet Jim on LinkedIn

06 / Come ready

Bring your curiosity.
Leave with a process.

Some familiarity with software development will help. You can follow the live demonstrations without installing an AI coding tool.

Join on Zoom. Register to receive your personal joining details. The full-day Claude Code, Codex, and AI coding workshop runs 9 AM–5 PM Eastern, including a one-hour lunch from 12:30–1:30 PM and Q&A throughout the sessions.

Code along if you choose. Bring a laptop, Git, an editor, and access to Claude Code and/or Codex. Any required AI tool subscription is separate from the free class.

Follow your organization’s rules. Use approved tools and synthetic training data. Keep government code, credentials, and sensitive information out of the exercises.

A few details

Good to know.

Is this really free?

Yes. Manicode Security is offering the full day of instruction at no charge. Optional hands-on use of commercial AI tools may require your own subscription or organization-provided access.

Do I need a government email address?

No. The class focuses on federal, state, and local government workers, political party organizations, and related public-service teams, but anyone is welcome to attend.

Is this affiliated with a political party or government agency?

This is an independent, nonpartisan class offered by Manicode Security. All U.S. political parties are welcome. No government or political party sponsorship or endorsement is implied.

What if my organization restricts AI tools?

You can watch the demonstrations without using Claude Code, Codex, or MCP yourself. Follow your organization’s policies and approval process before using any tool at work.

Can I ask questions or request accommodations?

We will use Zoom Q&A throughout the sessions. For accessibility requests or other questions, contact jim@manicode.com.

November 11, 2026 / Live on Zoom / $0

Better code.
Stronger public services.

A free day for the people who serve America.
Join Jim Manico, 9 AM–5 PM Eastern.

Register for the free class

Registration and joining details are handled through Zoom.