Project ideas from Hacker News discussions.

Big Fish: Exploiting a 37-Year-Old Kernel Bug in OpenBSD

📝 Discussion Summary (Click to expand)

Theme 1 – Skepticism about OpenBSD’s “heavily audited” security claims
Many commenters argue that the reputation of OpenBSD as exceptionally secure is overstated because the default install runs very little code and the auditing effort is limited.

  • “It’s remarkable that a bug survived 37 years in one of the most heavily audited codebases…” – ozereray1
  • “It's not as heavily audited as you think. They have like a dozen active developers.” – sippingabonedry
  • “The idea of 'OpenBSD is super duper secure' is mostly a fiction… Their prime claim… is pretty meaningless because the default install doesn't really do anything.” – stackghost

Theme 2 – Human‑ vs AI‑generated technical writing
The discussion splits over whether exploit write‑ups should be produced by humans (valued for clarity and insight) or can be supplemented/ replaced by AI (seen as either a cure for terseness or a source of low‑quality “slop”).

  • “Well, the key thing is that I don't want slop, period… if you expect a human audience to read about your findings, then is it not unreasonable to ask a human to write it?” – raddan
  • “If, in this era of constantly dealing with overly verbose verbal diarrhea from LLMs, your complaint is that a human write up is too terse… you can ask an LLM to turn it into as many pages of slop as you like.” – socializer
  • “…it appears that a human did write it, and then you complained that it wasn't good.” – john_strinlai

Theme 3 – The need for safer languages or verification to escape C’s pitfalls
Several participants point out that long‑lived bugs like this are symptomatic of using C, advocating for memory‑safe languages, formal verification, or augmenting C with safety features.

  • “I feel like this is the elephant in the room… C isn't a safe language and moving to safe languages is going to be the only lasting solution.” – zoeysmithe
  • “No language is safe… Alternatives are to formally verify the C code to be memory safe or to add some memory safety features to C.” – uecker
  • “Nobody in history has been able to write correct C code, and that includes OpenBSD developers.” – thomashabets2

🚀 Project Ideas

Generating project ideas…

VulnExplain – AI‑assisted security write‑up assistant

Summary

  • Helps researchers turn raw exploit data (crash logs, PoC code, diffs) into clear, structured bug reports with sections for root cause, impact, mitigations, and lessons learned.
  • Core value: reduces the time spent on writing while preserving human‑authored depth and clarity, addressing the frustration over terse or AI‑generated “slop” write‑ups.

Details

Key Value
Target Audience Security researchers, bug‑bounty hunters, teaching staff & students in security courses
Core Feature LLM‑guided drafting interface that takes inputs (binary, poc, logs) and outputs a markdown report with editable sections; includes citation of source lines and automatic CVE‑lookup
Tech Stack Python/FastAPI backend, open‑source LLM (e.g., Llama‑3) served via Ollama or HuggingFace, React‑based editor with ProseMirror, PostgreSQL for storage
Difficulty Medium
Monetization Revenue-ready: Subscription tiered by monthly report quota (e.g., $15/mo for 20 reports, $50/mo for unlimited)

Notes

  • HN commenter raddan said, “I don't want slop, period… a human should take the time and write it up,” highlighting the desire for quality over automated fluff; VulnExplain keeps the human in the loop while alleviating the writing burden.
  • Provides a practical tool for the educator who noted student write‑ups are “much better” than the terse AI summary, enabling consistent, high‑quality reports for grading and knowledge sharing.

KernFuzz Cloud – Continuous kernel/firmware fuzzing service

Summary

  • Offers on‑demand, coverage‑guided fuzzing of Linux/BSD kernels and embedded firmware images, automatically triaging crashes and estimating exploitability.
  • Core value: lets maintainers and device makers catch deep logic bugs (like the 37‑year‑old NetBSD bug) without needing large in‑house fuzzing infrastructure, addressing the worry that “fuzzing or static analysis could realistically have found this class of bug.”

Details

Key Value
Target Audience OS kernel developers, embedded device vendors, security teams auditing routers, cars, PLCs
Core Feature Upload a kernel/firmware image → service launches syzkaller/AFL++ instances in isolated VMs, collects crash corpora, runs lightweight exploit‑ability heuristics, and presents a dashboard with prioritized bugs
Tech Stack Kubernetes orchestrator, Go micro‑services, syzkaller & AFL++ binaries, Rust‑based crash analyzer, Prometheus + Grafana for metrics, React dashboard
Difficulty High
Monetization Revenue-ready: Pay‑per‑CPU‑hour (e.g., $0.10 per vCPU‑hour) with free tier for open‑source projects (up to 10 hrs/mo)

Notes

  • ozereray1 asked, “Do you think fuzzing or static analysis could realistically have found this class of bug?” – KernFuzz Cloud directly answers “yes” by making continuous fuzzing accessible and automated.
  • Enables discussion on HN about the prevalence of exploitable bugs in BSD/Linux and embedded systems, giving practitioners concrete data to move beyond anecdotal claims.

OpenSecScore – Transparent security‑claim evaluator for OS default installs

Summary

  • Scans an OS distribution’s default package set, identifies which services are actually enabled, and produces a clear “attack‑surface score” with CVE exposure, exploitability hints, and mitigation guidance.
  • Core value: counters misleading claims like “only two remote holes in the default install” by providing factual, data‑driven visibility into what a default install truly does, addressing concerns about inflated security reputations.

Details

Key Value
Target Audience System administrators, auditors, enterprises evaluating OS security, OpenBSD/Linux distro maintainers
Core Feature Input a distro release or ISO → tool enumerates installed packages, checks enabled unit files/services, cross‑references CVEs (NVD) and exploit databases, outputs a scorecard and markdown report
Tech Stack Python, ansible‑like package query (dpkg/rpm/apk), SQLite for CVE cache, FastAPI API, Svelte frontend, GitHub Actions for periodic re‑scans
Difficulty Low-Medium
Monetization Hobby (free, open‑source) – can accept donations or sponsor‑based hosting; optional premium API for bulk scans at $20/mo

Notes

  • stackghost criticized the “only two remote holes” claim as meaningless because the default install does little; OpenSecScore would give concrete evidence to support or refute such statements, satisfying the community’s desire for factual security discourse.
  • Encourages practical utility: admins can quickly see if enabling a particular daemon (e.g., httpd, smtp) introduces notable risk, informing hardening decisions without deep manual audit.

Read Later