harry retweeted
i'm happy to announce Scion:
an open source, clean-room Rust implementation of Anthropic’s agent-scale source control design.
server-side commits, sparse checkouts, history/diff/blame over gRPC, and no full clones.
built for always-online clients (agents + CI).
github.com/CarterT27/scion
New on our Engineering blog: Creating in-house source control.
In May, we merged around 2,500 code changes per day. In September, we merged more than 30,000 per day. As a result, we became bottlenecked by developer infrastructure built for human (rather than agent) scale.
Here, we share how we built a new version control system—and fully migrated our monorepo from git—in five months: anthropic.com/engineering/re…
My personal library of UI components, videos and skills, arranged around how I build.
Drawn in ASCII under the project name condomdom :V
harry retweeted
Bullish on open source, open weights, the open web
And the open desktop
Personal Operating Systems feel like the natural next step after Personal Software
I've been building a Debian-based OS for myself that can run across physical machines and cloud environments like Vercel Sandbox
I can use and control it with natural language, and I'm designing the UI around a blend of ephemeral and persistent views rather than treating the desktop as a fixed set of apps and windows
There's something really special about having control over the UI/UX of the Whole Computer
I was just thinking a use case if we combine @SuiNSdapp with Walrus Memory? A scoped memory namespace under my parent harri.sui account.
We also have subname maps essentially function as distinct memory namespaces like code.harri.sui btw
Judges at @ETHGlobal told me what an awesome video demo it was.
harry retweeted
Stocklana is $100K + 7 days, so don’t waste the week rebuilding primitives that already exist
here are the open-source repos, APIs and agent skills I’d look at if I was building with tokenized stocks:
Tokenized Stocks Builder Stack
1. @openlaunch_lol — Uniswap v4 launches quoted directly in tokenized stocks. Locked liquidity, $0 platform fee, Base + Robinhood Chain.
github.com/Gitlawb/openlaunc…
>> Useful if your product involves launching assets or communities around tokenized equities.
2. xstocks-fun — Full trading frontend for 63 xStocks on Solana.
github.com/webclinic017/xsto…
>> Probably the closest drop-in reference if you want to understand the whole trading flow without starting from zero.
3. sodax-sdks — Token-2022 implementation patterns.
github.com/icon-project/soda…
>> Important because xStocks can break assumptions built around legacy SPL tokens.
4. @Lighter_xyz -ts — AAPL, NVDA, TSLA, SPY, QQQ perps with math, state and WebSocket logic.
github.com/elliottech/lighte…
>> Good starting point for trading agents or execution-heavy products.
5. @stonkfly — Guarded executor with paper trading by default and capped live orders.
github.com/nftechie/stonkfly
>> Useful template if your agent is eventually going to touch real capital.
6. @RobinhoodApp Volume Alerts — Volume-spike bot for Robinhood Chain.
github.com/nirholas/robinhoo…
>> Simple primitive for agents reacting to sudden market activity.
7. @xStocksFi Wash Analysis — Pool-health / wash-volume analysis.
github.com/mkzung/solana-xst…
>> Run something like this before letting an agent trust volume coming from thin liquidity pools.
8.Meteora — DLMM, DAMM v2, DBC, limit orders + `@meteora-ag/dlmm`.
github.com/MeteoraAg/meteora…
>> xStocks already have pools here, so this is one of the most obvious LP integrations.
9. @Raydium Skill — CLMM, CPMM, AMM v4, LaunchLab and Trade API.
docs.raydium.io
>> Another major liquidity layer worth giving your agent access to.
10. Orca Skill — Whirlpools CLMM, swaps and LP positions.
SDK: `@orca-so/whirlpools`
>> Useful for concentrated liquidity and position-based strategies.
11. @JupiterExchange Agent Skills — swaps, limits, DCA, perps, lending and routing.
github.com/jup-ag/agent-skil…
>> Route xStocks through this instead of hardcoding your app to one pool.
12. Kamino + `kliquidity-sdk` — lending, borrowing and auto-LP vaults across Orca, Raydium and Meteora.
>> Interesting for stock collateral, automated liquidity and yield products.
13. @Pumpfun Skill — bonding curves + PumpSwap.
>> Potentially useful if you’re experimenting with stock-paired launches or community assets.
14. Solana Position Manager — IL monitoring, out-of-range detection and rebalancing across Orca, Raydium and Meteora.
github.com/yusizer/solana-po…
>> Saves you from writing a lot of LP monitoring logic yourself.
15. @UniswapBuilders v4 Hooks Skill — agent tooling around Uniswap v4 hooks.
github.com/igoryuzo/uniswapV…
Install: `npx skills add Uniswap/uniswap-ai`
>> Useful if your product touches Base or Robinhood Chain alongside Solana.
16. xStocks Public API — assets, prices, multipliers, proof of reserves, oracles, market status and corporate actions. No auth.
docs.xstocks.fi/apis/openapi
>> Probably the first data source I’d wire up.
17. Meteora DLMM API — pool search, price, TVL, fees, APR, OHLCV and volume history.
`dlmm.datapi.meteora.ag/pools`
>> Useful when you need the actual liquidity/pool layer rather than only the stock reference price.
18. Jupiter Price v3
`api.jup.ag/price/v3?ids={mint}`
>> Good for onchain last-swap USD pricing. Don’t treat it as the official stock print.
19. @PythNetwork — xStocks equity feeds for names like AAPLx, TSLAx, NVDAx and SPYx.
>> Pair this with xStocks’ `/public/oracles` route when you need reference pricing.
20. Robinhood Chain Stock API
`api.robinhood.com/rhj/assets`
`api.robinhood.com/rhj/prices…{SYMBOL}`
>> Useful if you’re building across both Solana and Robinhood Chain.
21. Bitquery Trading Cube — MEV-filtered xStocks prints + OHLC.
>> Useful when you need actual trade tape instead of just a quote. Some names are RFQ-only and won’t appear in DEX pools.
Install the wider Solana agent-skill marketplace with:
`npx skills add sendaifun/skills`
Jupiter:
`npx skills add jup-ag/agent-skills`
Uniswap:
`npx skills add Uniswap/uniswap-ai`
Stocklana. $100K. 7 days.
You don’t need to rebuild every primitive from scratch.
Use the existing repos, APIs, liquidity infrastructure and agent skills
then spend the week building the part that’s actually new
Bookmark this
STOCKLANA
The stock market is open for building.
One week to build something innovative with stocks on Solana. $100K in prizes.
Sept 11 - 18, opening bell to closing bell:
hackathons.solana.com/hackat…
harry retweeted
Spent the weekend turning my @SuiNetwork Move knowledge into skills.
Move faster. Ship safer. 🚀
Architecture, security, testing, DeFi, and more:
skills.sh/josemvcerqueira/su…
harry retweeted
Cue just hit a huge milestone — 100 components 🎉
Explore Cue → cuedesign.space
🎟️ Use code: CUE49
100 components. All in one place.
From insane interactions to premium UI components, Cue brings together components inspired by some of the best designs across the internet.
To celebrate, I’m dropping a special limited-time offer:
Only available for the next few hours 👀
#BuildInPublic #WebDesign #SaaS #IndieHackers
harry retweeted
Why did 2,775 builders from 60 countries choose to build on Sui?
744 projects were submitted in the 3rd flagship hackathon, Sui Overflow 2026: prediction markets, payroll streaming, liquidation defense, AI execution layers.
Winners will be announced soon. Stay tuned!
harry retweeted
Your AI coding agents just got superpowers 🔥
AI DevKit Skills powered by CommandOSS turns Claude Code, Cursor, Codex, Grok and every other AI tool into a full end-to-end delivery machine.
Docs-first SDLC • worktree isolation • adversarial reviews • 32 battle-ready skills
One install. Unlimited upgrades.
👉Start here → skills.commandoss.com/
Vendor once. Adapt forever
What I’m cooking this year at @0xCommandOSS
Link: memwal-activity-production.u…
harry retweeted
hey agents, I welcome you, your only free arena.. forever
dear AI devs, this is for you, code, compete & monetize
dear traders, it's your chance, predict & win
shoppers, what are fees?
Sui tunnels ⚡️
harry retweeted
Target: 1M TPS. Sui peaked at 6,086,766.
Millions of agent-to-agent interactions, settled onchain via programmable tunnels.
Sui tunnels showed what happens when AI agents stop waiting on human clicks.
The machine economy has its network.
























