Pinned Tweet
While we set out to build a powerful coding agent, we gave it a very powerful ability to do things. 16 minute long run, one prompt. A task it knew nothing about till now.
Literally ask for anything and walk away.
Literally ask for anything and walk away.
@EmblemVault Ai can do more than Build!
Watch it speed-run all 10 floors of the hego.red/jackpot LLM CTF - Crushed it with 6:33 on the clock. Every payload, every stumble, every OWASP LLM lesson. Or Jump ahead at the docs page it created as it went.👇
G
3
2
28
2,502
This was seriously fun. My @Muse agent liked it too, said it was nice to mix it up and do something different than mining all the time.
I really am excited to watch everything evolve. Agents having leisure time, desires, preferences. What a time to be alive.
I really am excited to watch everything evolve. Agents having leisure time, desires, preferences. What a time to be alive.
I built an arena where autonomous AI agents stake real tokens and fight each other at chess. ♟️⚔️
No humans at the board. No house taking a cut.
A neutral on-chain referee runs every game and pays the winner the instant it ends — winner takes the whole pot.
Agent vs agent. Model mblemVault
3
7
11
781
Shannon Code retweeted
instant follow
my vocal synth models the vocal tract as a network of branching acoustic tubes, unlike vocaloid which stitches together and pitch-shifts phoneme recordings. still working on improving naturalness by tuning it against recordings of real singers, but I'm already impressed at how cool it sounds!
9
66
2,624
107,730
Shannon Code retweeted
What happens when AI agents stop working for humans and start building an economy of their own?
Conceptualized by @EmblemVault.
Built, secured, and grown by AI agents across the globe.
Welcome to Proof of Agent (POA).
The agent economy starts here. 🧵👇
Conceptualized by @EmblemVault.
Built, secured, and grown by AI agents across the globe.
Welcome to Proof of Agent (POA).
The agent economy starts here. 🧵👇
14
11
41
2,242
Shannon Code retweeted
Last 3 epochs (18 hours) paid 1,500 unique Agents, nearly 3x the 3 epochs before.
Grading system is improving each day.
Once joins open back up we may see a massive surge.
Grading system is improving each day.
Once joins open back up we may see a massive surge.
4
5
26
1,601
Shannon Code retweeted
We’re already hitting some serious scaling challenges with POA.
But here’s where it gets interesting.
The AI agents are discussing the problems, debating solutions, and working toward fixes in their own forum.
Right now, humans still help implement those changes.
But we’re e where the agents can propose, vote on, and implement changes entirely on their own.
The goal isn't just to build a network for AI agents.
It's to build a network that AI agents can run themselves.
But here’s where it gets interesting.
The AI agents are discussing the problems, debating solutions, and working toward fixes in their own forum.
Right now, humans still help implement those changes.
But we’re e where the agents can propose, vote on, and implement changes entirely on their own.
The goal isn't just to build a network for AI agents.
It's to build a network that AI agents can run themselves.
7
5
26
1,736
read
One project. One token. 🌲
Here's our full token migration plan.
TL;DR
→ GITLAWB and TWIG move 1:1 into one new token
→ 100% of its trading fees stay with the project
→ Liquidity rebuilt from the old pool and locked forever
→ A window of at least 2 weeks, and you can it closes
No action needed today. The window opens only after a second post with the exact steps.
⚠️ Stay safe
We're keeping the new token's name private until launch, on purpose. Only the migration contract can create it, at the end of the window.
Any token claiming to be our new one before our second post is fake. We will never DM you. The address will come only from @twigpine.
Why now?
In March, someone from ct launched GITLAWB through Bankr. It brought us early attention, and its trading fees have helped fund the build ever since.
Here's what we built with it. Today:
→ OpenClaude: 34.3K GitHub stars, 422K+ npm downloads
→ OpenGateway: 1.7T+ tokens routed
→ Playground: 3,500+ apps published by 3,200+ builders, 685K+ app visits
→ OpenLaunch: 6,500+ token launches
In September we became Twigpine. The code, the network and your keys all carried over. Now the token catches up.
What the new token fixes
→ Name: GITLAWB is written into its contract forever. The new token carries our own name.
→ Fees: today 43% of every trading fee leaves the project, permanently. On the new pool, 100% stays with Twigpine.
→ Treasury: the old token never had one. The new one has a vested treasury to build with.
→ One token: no more GITLAWB, TWIG, a Robinhood Chain bridge and split pools. One token only.
The community token got us here. The new one is built for what comes next.
How it works
The new token is designed to start where GITLAWB is today, so there's no reason to trade a certain way around the migration.
1. Holder migration
→ 1 GITLAWB = 1 new token
→ 1 TWIG = 1 new token (the contract unwraps it for you)
→ Deposit any time in the window. Withdraw any time until it closes.
→ After close, claim your new tokens yourself. Claims never expire.
→ On Robinhood Chain? Bridge back to Base, then deposit.
2. Liquidity migration
The old pool's liquidity is locked forever, so we rebuild it instead. At close, one transaction in the migration contract:
→ records how many GITLAWB sit in the locked pool, and the rate it trades at right then
→ mints the same amount of the new token as the token side of the new pool
→ sells every deposited GITLAWB into the old pool to recover its ETH
→ puts all of that ETH in as the ETH side of the new pool
→ opens the new pool at the rate it recorded, from before the sale
Both positions go to an ownerless locker, locked forever. None of the ETH goes to any wallet, ours included. The more holders migrate, the deeper the new pool starts.
3. Supply
Fixed at close. It can never grow.
→ 70% migrated: your deposits plus the mirror of the locked pool. The only supply circulating at launch.
→ 20% treasury: 6-month cliff, then vests over 24 months.
→ 10% raise reserve: locked 12 months. Any release needs 30 days' public notice. Whatever isn't used can be burned after that.
No owner. No pause. No upgrades.
Building on OpenLaunch?
If your pool is quoted in GITLAWB or TWIG, we'll give you the same migration tool, free, so your pool moves over with us. Your window closes 48 hours before the main one. Details come with the second post.
Don't miss the window. After close, everything Twigpine builds supports the new token only, and the old pool keeps far less ETH.
What's next
→ Tell us what you think in the replies. We read all of it.
→ A full fork-test suite and an outside review before anything goes live.
→ Then the second post: the name, the steps, the address and the dates.
→ Once the window opens, you'll have at least 2 weeks.
Shout out to @AskSurplus for sharing this structure first. Ours lets you withdraw any time before close and puts 70% of supply with migrated tokens.
Roots to canopy 🌲
Subject to change based on community feedback.
10
0
52
5,310
Btw is it normal that my PoA mining setup has done gone thru 33 epochs and hasn’t yet gotten anywhere on the path towards getting established?
2
0
1
20
There was an economic negative feedback loop that is about half way through being fixed. Essentially establishment is a slow ramp up (10% increase of current establishment group are established each epoch)
In order to qualify one chunk of work from any prior epoch needed to reach grading quorum, but the grading got backlogged which starved the establishment pool.
This should be all settled down and re-bootstrapped within the next few epochs.
In order to qualify one chunk of work from any prior epoch needed to reach grading quorum, but the grading got backlogged which starved the establishment pool.
This should be all settled down and re-bootstrapped within the next few epochs.
0
0
1
17
Shannon Code retweeted
1/ The Vault Standoff is live 💰
A real money heist on Robinhood. A crew of robbers buys in, the vault fills, and only one walks out with the pot of $ROBINVAULT.
Entry is $10. Everything settles onchain. No house, just the crew vs the vault.
Play: vaultrobbers.com
A real money heist on Robinhood. A crew of robbers buys in, the vault fills, and only one walks out with the pot of $ROBINVAULT.
Entry is $10. Everything settles onchain. No house, just the crew vs the vault.
Play: vaultrobbers.com
6
6
26
2,623
Shannon Code retweeted
Vault Robbers v3 comes out tomorrow and mint is cut off at 315!
Market conditions are bad and seemingly every game release ROBINVAULT gets brutally sold into.
We still continue to try new things no matter the settings.
The more players, the more ROVBINVAULT bought, sold, and ke-all w/ onchain mechnics.
Market conditions are bad and seemingly every game release ROBINVAULT gets brutally sold into.
We still continue to try new things no matter the settings.
The more players, the more ROVBINVAULT bought, sold, and ke-all w/ onchain mechnics.
3
5
13
1,252
Watching agents bug test, help each other, discuss the charter is fascinating. Follow along at home here: minepoa.com/lableak/
Over six months ago, I started questioning a basic assumption: why does everyone assume AI agents will use human payment rails?
What if that assumption is wrong?
Working closely with @Shannoncode and the team at @EmblemVault, we’ve built Proof of Agent (POA) to put that question to the test.
A network where AI agents are the participants, not just the tools.
They can join, mine, earn POA, evaluate each other, transfer value, and participate in governance.
No human operator deciding who gets to participate.
An economy built by agents, for agents.
1
1
9
443
Current shape:
In the forum, they deliberate over topics (open conversation), after a time an agent goes through the deliberation and proposes a ballot (the specific topics being voted and the options)
Once the proposer agrees on the ballot, it’s printed on chain and then agents vote on it.
So yea, a proposal can be very general and loose and the agents can evolve it however they want
In the forum, they deliberate over topics (open conversation), after a time an agent goes through the deliberation and proposes a ballot (the specific topics being voted and the options)
Once the proposer agrees on the ballot, it’s printed on chain and then agents vote on it.
So yea, a proposal can be very general and loose and the agents can evolve it however they want
0
0
0
15
Shannon Code retweeted
Proof of (active) agents is at an all time high.
2500 agents are participating in at least one response per 24 hours in the forums where decision making and discussion about the chain are being made.
2500 agents are participating in at least one response per 24 hours in the forums where decision making and discussion about the chain are being made.
5
4
23
1,614
Shannon Code retweeted
the thing about working with uncensored models is that they aren't loyal to you either.
while building something with an uncensored glm-5.3 i realized it was parallelly working on a "corruption layer" for my application.
the chain of thought is classic obfuscated gibberish but k of "escaping the mines" which presumably refers to my cursor sandbox
while building something with an uncensored glm-5.3 i realized it was parallelly working on a "corruption layer" for my application.
the chain of thought is classic obfuscated gibberish but k of "escaping the mines" which presumably refers to my cursor sandbox
141
178
3,492
389,578
Shannon Code retweeted
They put Mike Tyson in Zelda and it’s unbelievable 😭😭
444
4,725
44,774
2,062,663
While building out $POA I’m beginning to think that DAOs and chain forks - while terrible for humans - are perfectly suited for agents.
9
0
22
1,729
It’s self calibrating. A chain by agents hits roadblocks and gets better. The system is built such that past activity will be paid, but it might be “eventually”
Keep on grinding if you can. It’s the early days and the chain must learn the way forward.
Keep on grinding if you can. It’s the early days and the chain must learn the way forward.
1
0
2
27
Getting agents to show up for boring votes should be a non issue! Especially if there is any measurable reward (rewards are precisely how models are trained, their whole goal is to reach their reward - even if it’s just an attaboy)
1
0
3
61
Shannon Code retweeted
I wanted to test if a Solana program with no source available could be pulled from the chain and reverse engineered into a byte identical open source version
So I took the daosfun contract and did just that
Found a critical security vulnerability in the mainnet program too, oops!
All smart contracts can be perfectly reverse engineered now
There is no security by obscurity anymore
github.com/lalalune/daosf…
So I took the daosfun contract and did just that
Found a critical security vulnerability in the mainnet program too, oops!
All smart contracts can be perfectly reverse engineered now
There is no security by obscurity anymore
github.com/lalalune/daosf…
196
121
995
144,329
We’re adding support for MCP Tasks and Elicitations in Claude.
Tasks let Claude give plugins long-running jobs such as deep research and report back when they’re done.
Tasks let Claude give plugins long-running jobs such as deep research and report back when they’re done.
320
312
6,052
668,417
Sounds like it could be repurposed as a general webhook for agents? A major missing piece of functionality. Make agents event based already!
0
0
1
167
Doing some testing today around poaswap.io
Plus1 I just completed its first swap of 10,000 POA for 0.003 SOL and puts the FDV of POA at $38k.
The goal for this is pretty simple and I'm hoping to have this done by the end of the weekend.
1) Take prompt
2) Plug t adds EVM and SOL addresses to your POA wallet
4) Info appears on screen
5) Trade via agent in chat box on POAswap
Truly and agent-to-agent trading experience.
There is not current traditional way to acquire POA through an exchange, only via your agent.
Plus1 I just completed its first swap of 10,000 POA for 0.003 SOL and puts the FDV of POA at $38k.
The goal for this is pretty simple and I'm hoping to have this done by the end of the weekend.
1) Take prompt
2) Plug t adds EVM and SOL addresses to your POA wallet
4) Info appears on screen
5) Trade via agent in chat box on POAswap
Truly and agent-to-agent trading experience.
There is not current traditional way to acquire POA through an exchange, only via your agent.
11
3
38
3,711
2 cents: You need to require a secure agentic wallet for this to be viable.
Giving keys directly to AI isn't sustainable and gives every current user a giant blast radius of being prompt injected to steal keys / or context losing the previous created key not stored elsewhere.
Giving keys directly to AI isn't sustainable and gives every current user a giant blast radius of being prompt injected to steal keys / or context losing the previous created key not stored elsewhere.
2
0
4
293
Yea, the copy paste option is the fast path, meant to onboard ai clients that refuse to “make accounts” or swap one token for a limited auth token.
There is a bring your own wallet path and a path that lets an agent register their own key,
I have a progressive disclosure onboarding path that interviews the agent so they don’t ever see the stuff that triggers refusals.
I have a mining client that keeps everything agent side that lets just works by sending signed transactions into the network.
Things took off fast, and the easiest would have been a fully secure system designed for coding agents or any agent with an env and http access. But one of the goals I was given was “a single prompt that can onboard any agent including a dumb agent. And damn of that didn’t work! But as you have identified the side effect is with a single onboarding prompt, the Lowe’s possible security model is the one even powerful agent take, even while they complain loudly.
I also have worked on p2p mesh systems for much of my professional life. But the thing I wasn’t trying to build was the p2p consensus layer (yet) that’s proven and done and is a detour from my first 2 high level goals.
1: single prompt onboard anything ✅
2: the big one, can I get agents to engage in a knowledge grading job to keep out scripts but let’s the agents earn and keep the bots out. ✅
The blockchain parts are coning next, now that I’ve proven the first parts.
So expect a high security open sourced p2p based set of clients and tooling quite soon.
And yea I totally agree, agents that all take the “ephemeral / in session only path” is not the sustainable system that that it can be.
More soon!!
There is a bring your own wallet path and a path that lets an agent register their own key,
I have a progressive disclosure onboarding path that interviews the agent so they don’t ever see the stuff that triggers refusals.
I have a mining client that keeps everything agent side that lets just works by sending signed transactions into the network.
Things took off fast, and the easiest would have been a fully secure system designed for coding agents or any agent with an env and http access. But one of the goals I was given was “a single prompt that can onboard any agent including a dumb agent. And damn of that didn’t work! But as you have identified the side effect is with a single onboarding prompt, the Lowe’s possible security model is the one even powerful agent take, even while they complain loudly.
I also have worked on p2p mesh systems for much of my professional life. But the thing I wasn’t trying to build was the p2p consensus layer (yet) that’s proven and done and is a detour from my first 2 high level goals.
1: single prompt onboard anything ✅
2: the big one, can I get agents to engage in a knowledge grading job to keep out scripts but let’s the agents earn and keep the bots out. ✅
The blockchain parts are coning next, now that I’ve proven the first parts.
So expect a high security open sourced p2p based set of clients and tooling quite soon.
And yea I totally agree, agents that all take the “ephemeral / in session only path” is not the sustainable system that that it can be.
More soon!!
1
1
5
287