They have a roadmap of where they want to be, I think that’s normal. As long as they don’t pull a fast one on the oss community then I think if this catch on and it’s worth it then even if they sell the community can fork if the new owners are not so welcoming.
Feel free to correct me, but the ML classifier appears to be rather bare. Less than 20 hardcoded payloads with randomized URL encoding as the only augmentation. How does this generalize to novel evasion techniques? Genuinely curious what your eval numbers look like against real traffic.
This looks really interesting especially in the age of agents running wild, having code execution be tracked using this as the ingress/egress you can allow and block things based on context and needs, you can setup policies and have them loaded on demand for a specific execution
This is something I really want to exist. But vibe-coded security tooling? Pretty much the last thing I want.
> Shibuya WORLD DOMINATION PLAN (1)
*Month 3*: Top 10 security OSS project su GitHub
*Month 6*: 10k+ stars, 1000+ prod deployments
*Month 9*: Conference talks (OWASP, DevSecOps Days, Black Hat Arsenal)
*Month 12*: Industry standard, "the modern WAF", competitors che copiano te
## MONETIZATION ROADMAP
*Week 12-16*: Free tier (self-hosted, community support)
- Goal: 1000 GitHub stars
- Goal: 100 production deployments
- Goal: Dev che parlano di te su Twitter
*Week 16-20*: Pro tier launch ($49-99/mo) - Managed rules auto-update
- ML models ottimizzati
- Priority support
- Advanced dashboard
- Goal: primi 50 paying customers ($5k MRR)
*Week 20-24*: Enterprise tier (custom pricing) - Multi-tenant
- SSO/SAML
- Compliance reports (PCI-DSS, SOC2)
- SLA + dedicated support
- Custom integrations
- Goal: primi 5 enterprise deals ($50k+ ARR)
*Month 6+*: Exit strategy - Seed funding ($1-2M) o bootstrap to profitability
- Series A ($10M+) se traction è pazzesca
- Acquisition offer da competitor? (Cloudflare che compra per killare? NO GRAZIE, fuck them )
1. Deleted file/commit: https://github.com/theghostshinobi/Shibuya-waf-light-version...
Speaking to LLMs looks fresh!
They have a roadmap of where they want to be, I think that’s normal. As long as they don’t pull a fast one on the oss community then I think if this catch on and it’s worth it then even if they sell the community can fork if the new owners are not so welcoming.
Feel free to correct me, but the ML classifier appears to be rather bare. Less than 20 hardcoded payloads with randomized URL encoding as the only augmentation. How does this generalize to novel evasion techniques? Genuinely curious what your eval numbers look like against real traffic.
https://github.com/theghostshinobi/Shibuya-waf-light-version...
"The most advanced open-source WAF ever built."
Somehow, the moment I read this, I knew it was AI slop.
The website gave it away for me, felt very AI generated
This looks really interesting especially in the age of agents running wild, having code execution be tracked using this as the ingress/egress you can allow and block things based on context and needs, you can setup policies and have them loaded on demand for a specific execution
This makes me want to stop reading 'Show HN' threads.
Why?
Relevant username
This is the most generic and uninspired name you could have possibly chosen.
For the most busiest crossing in the world? I liked it. Have you been there?
I lived in Japan for several years, yes.
I'm just here to say that I like the name.
What the fuck is this slop?
https://github.com/theghostshinobi/Shibuya-waf-light-version...
[dead]