Congratulations to the winners of the first week of the Skai Army sprints! 🥂

Your cybercats will be distributed shortly 🤝

Zealy usernames of the winners:
Parz
Orhaan22
Pawah
canitabrava7000
Gate_L.SKAI
Rypto_Coco_Jumbo
Fazquad
Diamondxy8
Alican123
OlSun

The second week of the Skai Army begins now! And with it a new chance of getting a cybercat 🔥

https://zealy.io/cw/skillfulai/questboard/0ef9a447-6dbd-4af0-abf5-2a2ead02a095/ebc004ce-4570-442e-8b7d-b0a91455605f

Cybercats have been sent 🙂

https://x.com/SkillfulAI/status/1861908803232600153

Updates pushed to production
- Agents Hub
- X/Twitter Skill (And with this skill credentials handling)
- Individual finance skills
- React Agent v1 (multiple tools usage)
- Big refactor around agent components (no visible impact though, but very useful for us to finish adding on the next layers of complexity)
- You should now see the agent hub populated with pre-made agents

Ecosystem
- New allowance system for model-hub & image-hub
- API keys enabled for all endpoints (main benefit for the Gaming SDK, but allows us to start moving towards letting external developers onto the platform)
- Pushed smart contracts on to SKALE Titan network for the badge system; Waiting on some front end resources to finish adding this to user profiles (Data is already being tracked)
- Referral Links - You can find this in your WALLET on the app. Currently these just get you more requests on the platform, in the future they will get you a cut of any payment the referred users spend on the platform

https://x.com/0xStinky_/status/1858305945362944151 - Create more vivid images (can also be done directly in chat with the image tool)

https://x.com/0xStinky_/status/1859760234916720774 - Connect twitter and have an agent tweet content for you.

Next update will include:
- Pump.fun token creation with the agent
(Reason we are not using $SKAI for pools based on ETH or BASE is because while short term it is super bullish, having the tokens tied up in volatile liquidity pools for agents is not really what it was designed for, not good if people lose interest. Instead we will take a cut of tokens created and use that to buy back $SKAI or send to staked TBD which option.)

- Custom SKAI React for better tool usage (Make the use of tools more intelligent and allow more complex tasks to be chained together)
- Image editing
- Task specific LLM Fine tuning for marketing feedback
- Image to Image models (Change one image into another ones style)
- Token/Agent minting
- Integrations server for telegram (Sets the foundation for lots of other external integrations also)
- Scheduling Agent tasks

Things are heating up 🔥



Live in production:

Pump.fun token creation with the agent
Custom SKAI React - Created Custom ReAct agent with improved prompting and skill handling
Image to Image models - we added image-to-image transformation in the image hub (Take and image, edit it's style - Currently making art pixelated)

Live in testing branches

Image editing --> We did half of this, put the rest on the back burner for partner needs.
Task specific LLM Fine tuning for marketing feedback --> This is a partner model and is almost ready to be given to them
Token/Agent minting - In testing, we have been really focused on the new locked staking contracts and auditing them
Integrations server for telegram - This was a huge infrastructure add, so we can tie it into scheduling and other apps outside of Telegram.
Scheduling Agent tasks - We got this onto our test net and have been able to schedule Tweets, working on adding scheduling to most other tasks at the moment

Other updates

Badge claiming/upgrading/showing refactor to use clean architecture
User entity creation along with use cases and migration of old functions to use new one
Added logic to the allowance/tier system so we can decide what features reset each month and which do not.
Refactor granting/removing/setting experience for users (Tied to badge claiming)
Cleaned a few lambdas that were no longer being used related to credits mostly
Other minor items refactored

Week 2 Cybercats have been distributed to:


Jessyuju
XGRACE
MrElkhorn.skai
Grinding
Bento555
Alican123
pacoporras90000
YosefO
Fazquad
Cardogonviral

Week 3 is now live and it will have twice the winners 🥂

Hey all,

I hope you are making the most of your festive season 🎄
It is important to take time to relax and reflect, and now is a good time as the market settles down for a few weeks after a historic run up.

2024 was a big year, and a lot happened. To some, we moved too slow; to others, they could not believe how much we got done on a comparatively tight budget.

There was a lot of uphill battles this year, but we are now in a position to move from tech focus building, to social expansion. This does not mean we are done building, not even close. It just means we are now ready to start the steps needed to push Skilllful out into the world.

## 1) Platform launch
To launch the platform in the way we want, we need several key components.
a) Smart contracts - We have been testing the basic version of these with our characters and marketplace, and expanding this to agents and beyond is on the horizon.
b) Fiat payments - We use the blockchain where it makes sense, and it is our main focus. But not implementing fiat solutions locks out a huge market. This has been a challenge, as a lot of payment providers reject anything that even has a whiff of cryptocurrency. Luckily after a lot of back and forth with dozens of providers, we are on the last steps to get them integrated.
c) Rewards - Everyone's favourite topic it seems. Post-launch, we will be doing rewards campaigns for the most impactful agents. This will be our pilot version of a rewards system. As the year goes on, and we add in more developer support, we will also be adding some funding rounds for builders using the platform. There will be more information on this as we go along.

## 2) Partner agents
We are currently working with several partners on fine-tuned agents with various capabilities. This will allow us to do several things
a) Show case - These partner agents will allow our infrastructure to be shown off in a functional way, one of them will also help aid in something you all ask more about. Marketing. We know this strategy can work well, as one of these agents was already built, and the team we built it for launched it. It is currently the most powerful autonomous agent we are aware of. We can not talk about that more for now, but all you need to find out more about it has been right in front of you all for several weeks.
b) Functional tools - While building these partner agents, we also make sure that any tool they made need e.g. blockchain explorer, is also designed in such a way, that we can add it to our public AI builder.
c) New partners - Partner agents indicates some kind of partnership, this means we can start adding partnerships in new areas. In the past we mainly focused on our infrastructure needs to get up and running. Now we can focus on what we can build for others.

## 3) Misc

a) Telegram - Telegram as a tool for speaking to agents has been something in the works for a few weeks. It took longer than other tools, as the infrastructure to use it, was built out to easily handle more credentials and apps in the future. This is currently in testing, and will allow people to create custom AI characters and agents that they can add to group chats. With extended capabilities as time goes on. These will be useful as QA bots to start, and will expand to handle more data as testing progresses

b) New agent info - We have overhauled the agent info screen, which will be live before the public release. It will be more informative, and is more like a product denoscription page. With more in-depth information, and more manageable sections for creators to handle their API keys, documents, and more.

c) Finally have access to xAI: GROK - In dev testing right now, we have integrated GROK. It was something people asked for a while back, but it was on lockdown. So you will soon be able to build agents powered by GROK.

d) Ask skillful - Added our documents to an agent that you can find on the app sidebar, where you can ask questions about our project. This will be updated as our documents are updated.

e) Scheduling, tools, and more - We are almost finished testing the ability to schedule your agents to do specific tasks. Allowing you to automate agent tweeting, data pulling, image generation, etc. We will work on making the task scheduling work with the majority of tools where this is useful, constantly expanding this feature. Image editing, short-videos, and more are all being worked on also. With access to googles models and infrastructure from their grant, these are all things they are helping us with integrating.

I also wanted to address something here, to do with new features. I have had a few people wonder how can it take us a few weeks to do 1 feature, if new projects pop regularly up with that feature for their agent (twitter was one, but 90% of those were just people manually tweeting). While I do understand some peoples frustration, we are not building 1 single agent that does XYZ, we have to build the framework, credential storage, access management, request tracking, etc etc, that allows thousands of people to create these agents. If we just build 1 agent, it would be easy to make it super impressive. As I mentioned earlier, we did this. We are making it so EVERYONE can build powerful autonomous agents, and add to an evolving ecosystem that becomes more powerful with every user.

2025 will be the year of expansion. Thank you all for your time, patience, and faith.

Sincerely,

Skillful AI

Reminder:

A few community members have flagged fake emails they are being sent from various emails.
Typically they are about claiming things.

We will never rush you on anything, always triple check things before opening or connecting your wallets. I should not have to say it, but it seems some people forget. Stay safe ❤️

Any new things would be through our official website only, and we would post about it here, X, and Discord.

We’re excited to announce that we will soon be migrating liquidity to improve trading conditions and enhance user experience! (YOU DO NOT NEED TO DO ANYTHING - We handle it all. Token stays the same, it is just upgrading the pools)

Going into 2025, our team will be creating a new Uniswap V3 (Base) pool and a Uniswap V3 (ETH) pool.

The new Uniswap Base (V3) pool will also be revitalized with USDC as the quote asset instead of USDT, offering enhanced flexibility and broader trading opportunities


This upgrade to V3 pools will allow us to take advantage of concentrated liquidity, offer more liquidity, better price stability, reduced slippage, and efficient trading for our community.

We’ll be migrating pools on Jan 2 (UTC), and will be announcing the official pool CAs here once we are done.

Stay tuned for more!

Hey, do not forget we are moving liquidity around today. If trading view, or any other app shows the liquidity dropped. That is why. When you move to a new pool it has a new address. So until the other apps update with the new data, they will show wonky things. That is why we made the post prepping you all.

The new contract addresses with liquidity will be shared in the next hour or so. You can likely find them already if you are in dire need to know them

We have now increased liquidity and moved to V3 pools. These are the new pool addressed. NOTE: Token address is the same, it is just the pool address that has been moved 🫡

Everyone is coming back from their quick breaks, and we are ready to move. A few people have found the agent we helped to create, and they seem as impressed as we are. We plan to do a lot of these in 2025 🔥

People have been impressed with chatbots on twitter, that is nothing compared to the specialisation we are building with others, and for you all on the main builder 🤖

https://dexscreener.com/base/0x0393cdffabdb750ab642e7ff9975247c58ff0fca

https://dexscreener.com/ethereum/0x2054df898bb8ccbf532ecf067291d994a8a9dc8a

pool address eth main: 0x2054Df898bB8CCbF532eCF067291D994a8a9Dc8A

Pool address Base: 0x0393CDfFABdB750Ab642e7ff9975247C58FF0fCa

Reminder to check pinned messages, getting a few messages about liquidity - As we said the other day, and as we shared recently, we moved the liquidity from V2 to V3 for better routing and better depth via concentrated liquidity.

None of you have to do anything, and you may also notice, you can buy $SKAI from $ETH or any other token, not just $USDT or $USDC (on base).

Stay safe, and we are working hard on the next steps.

X communty is ready: https://x.com/i/communities/1876767710815662169 🥂

UPDATE? Oh my sweet summer children, buckle up. I will lay down the long arm of the law on you all 🤝

1) Progress ≠ Linear
I needed time to build new partnerships and connections because, as soon as you stumble, people quickly lose interest in how good the technology is.

Progress isn’t always linear and obsolescence is always one cycle away.

That’s why it is important to find partners who see beyond the hype cycle, and who believes in systems rather than trends and slogans.


2) Why Did We Delayed Marketing?

Spending $20K−$50K monthly on pre-launch marketing was unsustainable.

Patience compounds capital.

In hindsight, it could have worked. But most of the action is on Solana for now, so even then, it might not have.

Fighting currents waste energy.

I believe Community is a by-product of asymmetric value…value that comes from creating real products that people want to use. And for that, full platform launch is a pre-requisite.

Hype is debt until the full product is launched.

3) Now, “The Plan”

So like a phoenix, here we are, preparing for a violent performance. We have saved the marketing budget the last few months, to release the platform swinging.

• Marketing: A six-figure budget for awareness, PR, and KOL partnerships.

• Partnerships: We have partners lined up. We have agents being built. We have a LIVE agent on X cleaning up the scum from the streets (people who follow me on X know).

• New React system: Integrating our React system (a framework of building blocks for creating websites or apps) enables our agents to perform multi-step tasks. It also allows us to not just use GPT models, but also use Google's models such as Gemini (For context, Google has provided us with hundreds of thousands of dollars in credits, which we've been using for testing over the past year).

• B2B Expansion: Our B2B arm is not limited to creating agents solely for clout, although we have done so in some cases. Moving forward, we will collaborate with [Redacted] and C[Redacted] to develop a wide range of personalized autonomous agents.

As you know, building an app for the average user to effectively use various AI models and the skills we develop to interact with them is significantly more complex than creating individual agents. However, with our framework, we are now well-positioned to launch and scale our B2B ventures.

• Skills: Once the platform is up, it is all about adding more Skills. Skills are expertise which you can give to your AI agent to do specialize tasks. We built the infrastructure for autonomous agents, and by our hand, we will build the use cases and skills too. We already have few skills on our platform like launching token on pumpfun and posting on X but many new skills such as [REDACTED] are coming soon..

More Token Utility
• Pay with $SKAI (Burning): You can subscribe to our monthly/yearly plan through credit card credit but we will allow users to unlock skills with $SKAI, that $SKAI is then burnt. You will also be able to pay for your agents’ functionality with card, or buy bundles with $SKAI.

• Holder Exclusives: Premium features reserved for $SKAI holders such as access to special skills.

What’s Next

• Skillful LLM: The works for this are already going on in the background [Redacted] LLM. Did you think we’d stop at skills and integrations? Not a chance. We're laying the foundation for custom fine-tuned models that specialize in specific areas. These models are designed to excel at specific tasks, and we're just beginning to scope out this potential after seeing success with smaller-scale models. Remember, models are just the brain, and companies will put out new ones all the time, including us.
But, but, but stinky, doesn’t training models cost millions?

Only if you want to build it from scratch and build a GENERAL model. We’re not.

Platforms like OpenAI and Anthropic battle for generality. We attack specificity (hey look, a new reasoning model just came out)

General AI requires scale. Specialized AI requires intent.
We’re building the latter.

Specialization > Size. A 5B model trained to specialize in one task (e.g., contract analysis, medical coding) outperforms bloated 100B models in speed, accuracy, and cost. This opens the door to running these models locally in the future (hint, hint). No cloud dependency, No latency and No privacy issues.

• Avatars: Details withheld. Expect disruption.

Our platform operates in a league of its own—unrivalled until you stretch the comparison to OpenAI or Anthropic. These are billion-dollar entities with near-limitless resources and years of head start.
That is what you need to be better.

Betting on us is betting on our system that outpaces the apex of the field, but without the inertia of scale (scaling for a start-up is easier than for a corporate giant).

In fact, the bet on us is so asymmetric you need to be careful not to fall over yourself.

Time bends only forward.

Infrastructure compounds.

And windows this glaring don’t stay open long.

Democratizing intelligence since 2025.