Two weeks ago we wrote Meet BugTraceAI — an open-source, self-hosted agentic pentester. The star count since then (365 at this moment) tells us people are watching. What we want now is conversation.
So we've enabled GitHub Discussions on the main repository, organised into three rooms:
- Feature Requests — tell us what the platform should do next. Nine threads are already open, from Nuclei integration to scheduled scans.
- Roadmap — weigh in on what we're building before we build it. Two live design questions: autonomous authentication/session handling, and an AI-powered HTTP repeater.
- General — secure defaults, deployment hardening, agent safety, anything else. The thread we most want answered: suggested secure defaults for local installs.
Why discussions instead of just issues? Issues are for bugs; discussions are for disagreements. A pentest tool embodies opinions — what to scan first, what counts as evidence, how much autonomy an agent should have — and opinions improve when people argue with them. We would rather hear "your default is wrong" in week one than discover it in a post-mortem.
Where to start if you have five minutes: AI report generation (what should an agent-written pentest report actually contain?), local LLM integration (air-gapped environments), or the API provider ecosystem proposal.
Every serious idea gets a response — usually within a day.
Repo: https://github.com/BugTraceAI/BugTraceAI
Top comments (0)