
AI Security Mesh
A zero-trust security system that monitors and controls actions performed by autonomous AI agents.
Gallery · Build with Mozaik
Every project submitted to the hackathon, live the moment it's submitted. Sign in with the email you applied with to leave a like or a comment.
Showing 40 of 40 · all tracks

A zero-trust security system that monitors and controls actions performed by autonomous AI agents.

A cross-channel scam detection system.

A multi-agent system for investigating production outages and recommending mitigations.

An incident-response system that investigates failures and requires human approval before applying fixes.

A multi-agent release review system that produces a ship-or-hold decision.

A multi-agent regression investigation tool that tests competing explanations against a running service.

A restaurant operations simulation where kitchen agents coordinate orders and disruptions.

A film-production scheduling system that resolves resource conflicts caused by unexpected disruptions.

A release-readiness system that verifies evidence before allowing a human deployment decision.

A content repurposing system that generates platform-specific content across multiple channels.

A multi-agent CLI for migrating software projects between frameworks.

An incident-response simulation where agents investigate competing theories and stop when evidence disproves them.
A food-rescue dispatch system that coordinates donations, recipients, and transportation.

A Windows desktop workspace where multiple AI agents collaborate on user requests.

A simulation that detects and prevents unsafe conflicts between concurrent agent decisions.

A platform for running game-theory and user-defined simulations with concurrent agents.

A non-destructive, multi-agent web security assessment platform.
A benchmarking tool that verifies whether a multi-agent system is genuinely concurrent.

An interactive heist simulation where one changed event causes autonomous characters to create a different story.

An incident-response system that rejects recovery plans when newer evidence makes them outdated.

A VS Code coding assistant that proposes changes through a human-reviewed diff.

An adaptive learning system where specialized agents research, teach, review, and assess a learner.

A multi-agent system for detecting scams and phishing attacks.
A multi-agent incident war room with safety guardrails and human approval for production actions.

A tool that investigates whether a public GitHub bounty is funded, available, and eligible.

A live incident command room where agents investigate outages, propose fixes, and block dangerous commands.

An automated system for investigating outages, proposing mitigations, and drafting status updates.

A multi-agent system that searches for real-money opportunities requiring no starting capital.

An automated cybersecurity system that analyzes attacks, applies WAF mitigations, and checks compliance.
A simulated VC panel that evaluates startup pitches from growth, technical, and risk perspectives.

A live arena where multiple language models answer and challenge one another.

A bill-processing system that requires approval before payments or cancellations.
A multi-agent moderation system for detecting spam, invalid posts, and harmful content on a Nostr relay.
A coding CLI where one agent fixes a bug while another independently reviews the solution.

A multi-agent research tool that produces sourced and verified answers.

A developer workspace for researching, reviewing, comparing, and troubleshooting code with specialized agent teams.
A multi-agent system that investigates claims from supporting, opposing, and contextual perspectives.
A tool that verifies the availability, payment terms, and eligibility requirements of public software bounties.

A multi-agent incident command system that investigates failures and proposes safety-checked recovery actions.

A security operations system that monitors Solana protocols and requests human approval for protective actions.