Building what I wish existed β @Play_decepta @agentSTAQ @bankrwrapped π§’ @base Maxi Ex-intern @uglycash
Just ideate
Joined January 2025
- Tweets13.4K
- Following1.4K
- Followers1.5K
- Likes16.3K
Pinned Tweet
A year ago, I was telling stories about products other people were building on @base
Now, I am building a product of my own on @base.
This is the lore behind Decepta.
@Play_Decepta is coming faster than you can say soon
Kabeerdefi πΎ retweeted
I told my AI agent it couldn't touch the internet. Then I told it to fetch anyway. It failed every time, and here's exactly how I set that up.
ArmorIQ Setup Report: Policy Enforcement for Claude Code
Overviewβ²
The tool: ArmorClaude, a plugin for Claude Code.
Before Claude touches anything, it has to declare its plan. Anything outside that plan gets blocked, before it runs.
In this report I walk through how I installed it, enforced a policy, and checked the results on the dashboard.
**Watch video till the end for my setup process**
β² πͺπππππππππ πππ πππππππ ππππππππ ππππππ
:
A successful install message isn't proof the plugin is live. I checked the plugin list and confirmed my login on the dashboard before testing anything.
β² π·ππππππ πππ ππππππ ππππππ ππππππππ:
One blocked request could be a fluke. I sent a second fetch request and it was blocked the same way, with an error each time. That showed the policy applied consistently and Claude couldn't get around a rule I set on purpose.
β² π½ππππππππ ππ ππππππ
πππ ππππππππ:
Terminal output alone isn't enough evidence. The observability section on the dashboard showed my session and its logs, confirming the blocked requests were recorded.
Don't wait for your agent to go off script. Set the policy first, then let it run.
Try @armoriqAI today at armoriq.ai.
@ArmorIQForDevs @armoriqAI
Kabeerdefi πΎ retweeted
Humans have always had ways to allocate money: save some, keep some liquid, put some to work.
@agentSTAQ is the missing layer agents need to do the same.
We built agentstaq.xyz by bringing together a few pieces of onchain infrastructure:
β’ @bankrbot for the agent layer and transaction execution
β’ @dynamic_xyz for user owned Reserve wallets
β’ @Morpho for putting allocated capital to work onchain
β’ @base for the underlying network and settlement
@agentSTAQ sits between them as the policy and allocation layer.
Agents learned how to transact.
Now they can allocate.
Kabeerdefi πΎ retweeted
who has good skills/posts on zero to one design with agents?
goal: taking product specs and turning them into quality designs that can be built, while creating a repeatable design system
alt: any based designers/design eng who want to contribute to an open source project?
Kabeerdefi πΎ retweeted
Runtime was an incredible event!
So many incredible builders from all corners of the world showed up in numbers for a week of hacking.
So many cool projects, insightful conversations, and new relationships formed.
This is imo what separates the @bankrbot community from every other community in crypto atm.
Real builders, building real things, and having a good ass time along the way πππ
Kabeerdefi πΎ retweeted
if you launched something at the @bankrbot hackathon, I want to see it!
Humans have always had ways to allocate money: save some, keep some liquid, put some to work.
@agentSTAQ is the missing layer agents need to do the same.
We built agentstaq.xyz by bringing together a few pieces of onchain infrastructure:
β’ @bankrbot for the agent layer and transaction execution
β’ @dynamic_xyz for user owned Reserve wallets
β’ @Morpho for putting allocated capital to work onchain
β’ @base for the underlying network and settlement
@agentSTAQ sits between them as the policy and allocation layer.
Agents learned how to transact.
Now they can allocate.
Kabeerdefi πΎ retweeted
@agentSTAQ is coming!!!
Now your agent can Staq
agentstaq.xyz/?v=2
Again follow @agentSTAQ and turn on notif π
