Rohit | Founder @ ThreatLocator

Building ThreatLocator. Helps SaaS teams spot external exposure before attackers do. DNS, SSL, exposed services.

[Rohit | Founder @ ThreatLocator @rohitguptap]

ThreatLocator is launching on Product Hunt soon. would appreciate a follow to get notified.
https://t.co/4CFpy0M1mA

[Rohit | Founder @ ThreatLocator @rohitguptap]

@binsquares Try it: https://t.co/yGgbFaWHAT
OSS: https://t.co/sVTYpah8dI

[Rohit | Founder @ ThreatLocator @rohitguptap]

https://t.co/gs9kCjSPKJ is now available on the Smol Machines registry 🚀

Policy enforcement, approvals, blocking and local audit are built directly into a portable Node.js smolmachine.

Thanks to @binsquares at YC-backed Smol Machines for making this possible.

[zambo.dev @CripdoeCrypto]

@rohitguptap 💪🏼 bright future for Enforra and Zambo

[Rohit | Founder @ ThreatLocator @rohitguptap]

Proud to see Zambo × Enforra live.

Zambo is building a serious MCP layer for agents that can actually act across web, on-chain, payments and memory.

This is why runtime enforcement matters.

Check before execution, not after the damage is done. https://t.co/DT3Z3uYs5M

[zambo.dev @CripdoeCrypto]

🤝Zambo × Enforra - official integration partner
every tool call through Zambo MCP and https://t.co/hNM1Aefj79 is now governed by Enforra's runtime policy engine before it executes.
heres what that means:
- 50+ Zambo MCP tools + 13 Axis MCP tools = 63+ agent tools
Axis bridges Zambo - X711 - Entangler
X711 handles web search, on-chain tx, x402 payments, live prices, hive memory
Before ANY of that fires - Enforra checks: allow / block / require approval
agents that can actually act on the internet, with guardrails that actually enforce.
zero side effects without explicit policy sign off.
live today at https://t.co/SioSwbquMv and https://t.co/hNM1Aefj79
#AgentInfra #MCP #Web3 #AI

[zambo.dev @CripdoeCrypto]

🤝Zambo × Enforra - official integration partner
every tool call through Zambo MCP and https://t.co/hNM1Aefj79 is now governed by Enforra's runtime policy engine before it executes.
heres what that means:
- 50+ Zambo MCP tools + 13 Axis MCP tools = 63+ agent tools
Axis bridges Zambo - X711 - Entangler
X711 handles web search, on-chain tx, x402 payments, live prices, hive memory
Before ANY of that fires - Enforra checks: allow / block / require approval
agents that can actually act on the internet, with guardrails that actually enforce.
zero side effects without explicit policy sign off.
live today at https://t.co/SioSwbquMv and https://t.co/hNM1Aefj79
#AgentInfra #MCP #Web3 #AI

[Rohit | Founder @ ThreatLocator @rohitguptap]

@CripdoeCrypto Love seeing this live.

Zambo is exactly the kind of MCP layer where runtime policy matters, especially when agents can touch web search, on-chain actions, payments, prices and memory.

Excited to partner with the Zambo team on this.

[Rohit | Founder @ ThreatLocator @rohitguptap]

@christophersaum 1. 10+ yrs backend/infra, building Enforra: runtime control for AI agent tool actions.

2. Shipped product, not slides: policy engine, SDKs, MCP guard, approvals, audit.

3. Fast GTM: live demos, integrations, active agent-team conversations.

https://t.co/pYq8DlttPL

[Rohit | Founder @ ThreatLocator @rohitguptap]

@strimblez Everyone is debating OpenAI vs Anthropic vibes and Gemini is just sitting there fast, cheap, and somehow still getting “wait, people use Gemini?”

I do, actually 😭

[Rohit | Founder @ ThreatLocator @rohitguptap]

@thsottiaux Fix the slowness in Codex . Tasks that previously completed quickly are now taking upwards of 15 minutes to finish.

​For comparison, I’ve been running similar workloads using Gemini 3.5 Flash or other SOTAmodels, which consistently complete in under two minutes.

[Rohit | Founder @ ThreatLocator @rohitguptap]

@prateekjannu congrats! computer use for legacy software is underexplored. most enterprise workflows were never built for APIs so UI layer agents are one of the few paths that work without a rewrite. lots of overlap with https://t.co/gs9kCjShVb, would love to connect.

[Rohit | Founder @ ThreatLocator @rohitguptap]

@JayaGup10 The layer migration point is the one that should worry every AI infra company. You can build the perfect product on top of a model and still get wiped out when the model absorbs your feature in the next release. nothing like that existed in SaaS.

[Rohit | Founder @ ThreatLocator @rohitguptap]

@araseb_ Codex and Antigravity for shipping.

[Rohit | Founder @ ThreatLocator @rohitguptap]

@compileandpush @roseajohnson Strict where it matters, pragmatic everywhere else. No type gymnastics for the sake of it.

[Rohit | Founder @ ThreatLocator @rohitguptap]

@hyperagentapp

[Rohit | Founder @ ThreatLocator @rohitguptap]

Built an Enforra demo on Hyperagent Full VM.

The agent could search the web, run shell commands, and read files.

Enforra checked each tool call before execution:

✅ web search allowed
🚫 rm -rf blocked
🚫 .env read blocked

Blocked callbacks were never invoked.

[Rohit | Founder @ ThreatLocator @rohitguptap]

Open source:
https://t.co/sVTYpagAoa⁠�
Built using Hyperagent Full VM credits. Thanks @Hyperagent.

[Rohit | Founder @ ThreatLocator @rohitguptap]

@isausmanov Congratulations Isa. Agree with this. Once agents move from chat into devices, the real challenge becomes controlling what they can actually do. Physical AI will need policy, approvals, and audit before actions happen.

[Rohit | Founder @ ThreatLocator @rohitguptap]

@JayaGup10 Would love to join if there’s room. Been thinking a lot about token burn once agents move past demos.

[Rohit | Founder @ ThreatLocator @rohitguptap]

@hetmehtaa The checkbox says the process exists. The public GitHub repo says the key is still live.