Iftakhar Rahmany

AI agents are starting to touch money, credentials, commerce, and real life.
Building 1ly + DCP: commerce and permissions for that world.

[Iftakhar Rahmany @theiftakhar]

https://t.co/YNTcBs2L4n

[Iftakhar Rahmany @theiftakhar]

this move shows how little the government understands digital platforms. Millions of legitimate users depend on Telegram for business, education, and communication.
Punishing everyone because some criminals misuse the app is not smart governance.

[Pavel Durov @durov]

India’s IT ministry banned Telegram for one week because some users shared leaked exam questions.

This punishes 150M+ ordinary Telegram users in India — not the insiders who leaked the exam materials.

And the ban hasn't stopped anything. The leaks just moved to other apps.

[DCP @dcpagent]

DCP hit some wild numbers in the last 30 days:

2541 npm downloads
93 desktop installs
108 Product Hunt upvotes
250+ developers cloning code

But stats don't tell me what I really need to know.

How are you using it? What do you love? What needs fixing?

Join discord 👇

[Iftakhar Rahmany @theiftakhar]

@productwatch_io @betterdbhq Thanks for the shoutout
@dcpagent

[Iftakhar Rahmany @theiftakhar]

@0xBenii @solana @dcpagent

[Iftakhar Rahmany @theiftakhar]

DCP should be on this map.

We’re building the permission wallet layer for x402 agent payments.

Agents can pay for APIs/services, but users set the rules:

daily spend limits
merchant/API allowlists
auto-approve small payments
manual approval for larger ones
revoke anytime
audit every payment

x402 handles payment.
DCP handles user-controlled authority.
@dcpagent

[Iftakhar Rahmany @theiftakhar]

folks losing their minds about Hermes Agent going goblin mode and throwing random errors like `NoneType`

just open terminal and run:

`hermes update`

It’ll work after that, thanks me later.

@NousResearch

[Iftakhar Rahmany @theiftakhar]

@bradmillscan It’s not entirely Hermes, problem was with Chatgpt end. They updated it so just update your Hermes and it should work like a magic

[Iftakhar Rahmany @theiftakhar]

Awaek is on OpenClaw now :)
openclaw skills install awaek

[Iftakhar Rahmany @theiftakhar]

Awaek hit 110+ GitHub clones + 10 stars in under 40 hours
Didn’t expect this much early interest. Thank you everyone testing it and sending feedback.

Next up: OpenClaw support.

Your saved X bookmarks are becoming a personal source engine for what to build, post, launch, and research next.

https://t.co/X1wRWjrV6v

[Iftakhar Rahmany @theiftakhar]

@steipete Hmm, sounds similar to what I was building. May be I’ll add birdclaw support as well

https://t.co/c71YZQUWpm

[Iftakhar Rahmany @theiftakhar]

Safest approach for sure. But let’s be honest, most of us get lazy sometimes, and hidden files/prompts are easy to miss.

That’s why I think agents need to be safe regardless of what skill or tool they use.

I’m building DCP for this: connect any agent with one command, then explicitly define what it’s allowed to do. Sensitive actions like spending money, accessing API keys, or touching private data require permission first.

Agents should be useful for real work, without giving every skill unlimited power.
https://t.co/UXn86n5SUN

[Iftakhar Rahmany @theiftakhar]

This is why every agent needs a permission layer like DCP.

Agent skills, plugins, third-party tools, and hidden prompts can be malicious or compromised. Once your agent has access to money, API keys, accounts, or private data, a bad skill can try to misuse them.

DCP protects the user by making sure the agent can only do what it was explicitly allowed to do.

So even if a tool tries something shady, it can’t drain your wallet, leak your keys, access private data, or take sensitive actions without permission. The agent gets autonomy, but not unlimited power.

https://t.co/Fc0JD8ZBbK

[Iftakhar Rahmany @theiftakhar]

This is why every agent needs a permission layer like DCP.

Agent skills, plugins, third-party tools, and hidden prompts can be malicious or compromised. Once your agent has access to money, API keys, accounts, or private data, a bad skill can try to misuse them.

DCP protects the user by making sure the agent can only do what it was explicitly allowed to do.

So even if a tool tries something shady, it can’t drain your wallet, leak your keys, access private data, or take sensitive actions without permission. The agent gets autonomy, but not unlimited power.

https://t.co/Fc0JD8ZBbK

[Iftakhar Rahmany @theiftakhar]

Why is it that @claudeai or @ChatGPTapp respond better on app?
Same model same llm but different responses? It’s me only or someone else feels same?

[Iftakhar Rahmany @theiftakhar]

@barkmeta Let’s revive NFT, I got insane idea

[Iftakhar Rahmany @theiftakhar]

@barkmeta @barkmeta check your DM pls ( definitely not collab promo and spam related generic messages). Commenting here so you find my msg

[Iftakhar Rahmany @theiftakhar]

@koharishant Congrats Ishant, CASE looks really interesting.

I like the direction: agents become much more useful when they can actually operate real apps, not just generate instructions.
upvoted 🙂 good luck with the launch.

[Iftakhar Rahmany @theiftakhar]

I’m launching on Product Hunt for the first time tomorrow and honestly feeling a bit nervous.

Not sure if I did the page right, the positioning right, or even the launch setup right.

I’ve built a lot of things before, but PH feels different.
product is DCP: a local permission vault for AI agents.

Simple idea: agents should not hold your private keys, API keys, or sensitive data.
They should ask for permission.

I scheduled the launch for Alpha Day & preview is live now.

Would genuinely love feedback from people who understand PH.

What did I do right?
What feels unclear?
What would you change before it goes live?

Preview:
https://t.co/mYJYR2XDeb

@ProductHunt

[Iftakhar Rahmany @theiftakhar]

@shifthappends @AlexFinn haha now ask Awaek about Awaek

[Iftakhar Rahmany @theiftakhar]

https://t.co/bzR3gJbXu8

[Iftakhar Rahmany @theiftakhar]

The next internet user is not human.
It’s an AI agent with:
- tools
- API access
- wallets
- budgets
- jobs to finish

But we’re still giving agents secrets like this:

“here’s my API key”
“use my private key”
“please don’t mess up”

That model breaks fast.

Agents should ask for permission, not hold secrets.

Launched DCP on @ProductHunt today to make that real.

Feedback welcome: