adshao/flounder

Autonomous white-hat security auditor for AI-driven code review, bug bounty research, exploit construction, and execution-grounded verification.

What it solves

Flounder is an autonomous white-hat security auditor that automates the end-to-end process of auditing software targets. It replaces the manual, tool-switching workflow of security researchers by providing a structured system for target preparation, attack surface mapping, deep-dive auditing, exploit construction, and verification of findings against real-world ground truth.

How it works

Flounder acts as a thin workflow layer around a coding LLM. Instead of relying on hard-coded bug rules or scanners, it allows the model to decide the audit strategy while providing the necessary infrastructure:

  • Autonomous Loop: It executes a pipeline of prepare $\rightarrow$ map $\rightarrow$ dig $\rightarrow$ synthesize $\rightarrow$ verify $\rightarrow$ confirm $\rightarrow$ report.
  • Sandboxed Execution: Model-generated PoCs and tests run in an isolated OCI backend to prevent host pollution and reduce the blast radius of malicious dependencies.
  • Execution-Grounded Proof: A finding is only upgraded to "confirmed" if the model can provide a passing local command that exercises the vulnerable path.
  • Target Preparation: It can automatically turn a repository link, transaction hash, or address into staged audit materials.

Who it’s for

It is designed for white-hat security researchers, bug bounty hunters, and developers performing authorized security reviews, particularly those working with Solidity/EVM and ZK/proof-systems where local forks and harnesses can prove vulnerabilities.

Highlights

  • Framework-Agnostic: The audit strategy is driven by the model, meaning it supports any stack (Rust, Go, JS, etc.) as the model improves.
  • Network-Sealed Discovery: Discovery runs are network-sealed to ensure findings are derived from the target material rather than copied from existing disclosures.
  • Durable Evaluation: Supports running validated positive/negative control groups to measure audit performance.
  • Agent-Ready: Integrates with coding agents like Codex and Claude Code via a skill system, and provides a React dashboard and REST API.

Related

  • Project
    lintsinghua/DeepAudit

    DeepAudit is an open‑source, multi‑agent AI security auditor that analyses code, generates and sandbox‑executes PoC exploits, and outputs professional reports. It runs via Docker, supports local LLMs, and aims to reduce false positives and verify vulnerabilities automatically.

  • Project
    StanleyNull/AutoHunter

    AutoHunter is an AI‑driven, Docker‑based platform that continuously hunts for web vulnerabilities. It uses LLM‑guided agents (Collector, Worker, Reviewer) together with real tools like nmap, nuclei, and sqlmap. Humans only perform final review. The system supports multiple search engines (FOFA, Quake, Shodan, etc.), offers a Vue dashboard, and persists data in SQLite. It’s licensed Apache 2.0 and intended for authorized security testing.

  • Project
    vndee/llm-sandbox

    LLM Sandbox is a Python package that runs code generated by large language models inside isolated containers (Docker, Kubernetes, or Podman). It supports multiple languages, enforces security policies, can capture output plots, offers interactive notebook‑style sessions, and includes a container‑pooling system for fast, concurrent execution.

  • Dispatch
    Cloudflare security-audit-skill: Open‑source multi‑phase security audit framework for coding agents

    Cloudflare’s security-audit-skill adds a six‑phase, verifiable security audit workflow to coding agents, enabling reproducible vulnerability discovery with independent validation and structured reporting.

  • Project
    Kritt-ai/open-kritt

    open·kritt is an open‑source, self‑hosted platform that orchestrates AI agents (via Codex, Claude Code, Grok Build, etc.) to run custom, prompt‑driven security workflows on code repositories. It builds focused tasks, runs them in parallel Docker containers, validates findings, de‑duplicates and ranks them, and exports a complete report bundle. Designed for bug‑bounty researchers and DevSecOps, it offers a web UI and a headless CLI, supports any model provider, and is released under AGPL‑3.0.