Show HN: Vigil – Open-source AI pentesting and code audits

Category: security

Tags: ai-security, pentesting, code-audit, llm-agents, bug-bounty

Score: 7.5/10 (Innovation: 7, Technical: 8, Documentation: 8, Utility: 7)

Vigil is an AI-driven harness for code audits and web pentesting that coordinates LLM agents (like Claude Code) through multi-pass investigation, runtime reproduction, and evidence tracking. Its methodology—including specialist agents, verification challenges, and session handoffs—was refined through real bug bounty work, setting it apart from simple LLM wrappers.

Target audience: security engineers, bug bounty hunters, application security teams

Repository: https://github.com/VigilOSS/Vigil · Python · Apache-2.0 · 29 stars

View on Hacker News