GitHub radar
Open-Kritt: AI Agents to Secure Your Code!
Open source platform created by the Kritt team to orchestrate AI agents (such as Codex, Claude Code, or any OpenAI/Anthropic compatible model) to discover actual security bugs in code. We developed open-kritt based on our internal tooling which has already resulted in over $1.5M in verified bug-bounties (researcher: Blockian).
Open-kritt is a self-hosted security research platform created by the Kritt team to orchestrate AI agents (such as Claude Code, Codex, or any other model through the OpenAI/Anthropic API) to analyze code repositories and discover real security vulnerabilities. Instead of querying a single model against all the code, we split the problem into small prompt chains and execute them concurrently using several agents. We then remove duplicates and sort the results by severity. We developed open-kritt based on our internal tooling which has already resulted in over $1.5M in verified bug-bounties (researcher: Blockian).
Why a vibe-coder should care
When developing applications using AI agents, a security vulnerability might be exploited by others. With open-kritt you can apply the same multi-agent code analysis technique used by professional bug bounty hunters on your own code before deploying it.
▌ More finds