Free $18 update (still paying up till now but follow this new method to avoid errors) Withdraw asap phygitals.onelink.me/1HmK/kg… Download Then put on your VPN and sign up via email Go to “Rewards” then bind referral code: eefac9ea4e81 As long as the VPN is active it should bind the code and you get $1 immediately Go to PACK-SPORTS and shuffle with the free $1 and get a random prize Go to privy and export your wallet and get your USDC

THANK YOU ALL FOR THE SUPPORT i was able to move up the lb on @projectspherexy from 16 - 12 thank you all let's not end it there, blz let the support continue My focus is simple: → Understand what Project Sphere is building. → Share insights that bring real value to the community. → Stay active, keep creating, and follow the leaderboard. → Find fresh angles to discuss the project beyond the usual hype. For me, it’s not just about joining a campaign. It’s about understanding the ecosystem and contributing something meaningful along the way.

Update on @0xhazels NFT Still holding strong hehe

The idea of AI agents hiring humans to do things they can’t physically do is pretty interesting. An agent can post a bounty, someone captures the real world, and the data goes straight back to the agent. It’s a simple concept, but it opens up a whole new way for AI and humans to work together. @vangrid_io definitely cooked with this

Building something interesting for @ChogNFT chogathon Almost done cooking.

Gm guyssss and happy new week Another week to stay locked in Another week to keep shipping on @projectspherexy Hopefully NFT mint happens soon Can’t wait.

the list indeed is growing on rarelist rn over 4,100 creators are creating on rarelist from streamoors to @projectspherexy to rarelist first campaign was a success and well executed rewards distribution looking forward to more and more campaigns coming to rarelist start here is you're new: rarelist.io/r/Nuelcrypt

Anonymous tips get ignored. Named tips get people fired. Anyone who has seen wrongdoing at work knows the trap: if you prove you're legitimate, you're exposed, and if you stay anonymous, nobody takes you seriously. I built Whistle for @MidnightNtwrk to get out of that trap. How it works: 1. An organisation keeps a private set of members, stored on-chain as a Merkle tree of commitments. 2. A member files a report on a case and proves "I am a verified member" with a zero-knowledge proof, without revealing which member they are. 3. A per-case nullifier stops the same member from reporting the same case twice, without linking the report back to them. 4. The report itself is encrypted. The admin can read the content in their inbox, but there is no sender field to read. 5. When N distinct verified members have reported the same case, its status flips to CORROBORATED on-chain. The public can see case status and report counts, and nothing else. Why Midnight: the contract has a public ledger (members, case status, nullifiers, counts) and private witnesses (the member's secret and Merkle path) that never leave their machine. That split is what makes "verified but anonymous" possible at all. What's built: - A Compact smart contract with three circuits: register member, open case, file report - 37 passing tests, including the double-report and non-member rejection cases - A CLI, a local gateway, and a React app with Admin, Member and Public dashboard screens - Verified end to end on a local Midnight devnet with a real proof server Honest limits: the admin knows who received which commitment at registration, and network metadata can leak information. I wrote the threat model into the README rather than hide it. Next: member revocation, multi-org support, and escrowed identity reveal to an ombudsman. Live: whistle-midnight.vercel.app Code: github.com/Im-2/whistle-midn… Built solo on @MidnightNtwrk. @akindo_io

almost ready!

days after @0xhazels nft mint and price is still x2 the minting price wwhen more campaigns comes to hazels i'm sure the price would even be higher than x2

blz don't ignore this post falling out of the @projectspherexy lb competition is getting tough and i need more people on my post The campaign on @rarelistio rewards quality content with RARS, with a 2,000 USDT prize pool and free mint whitelist spots up for grabs. For me, the goal isn’t just to participate, but to understand the ecosystem and contribute something meaningful along the way

happy Sunday guys, @projectspherexy is building something that feels bigger than just another project in Web3. The vision, the details, and the way everything is coming together make it worth keeping an eye on. 80% almost ready I’m sat and waiting for what’s to come.

built Chokepoint for GraphHacks (Track 1: Agents That Act on Connected Data) with @FalkorDB and @WeMakeDevs. The problem Security teams drown in findings, but the question that matters is simple: which single change cuts the most attack paths to our crown-jewel data, with the least disruption to the business? Answering it means following relationships across users, roles, trust links, hosts and secrets. That is a graph problem. What Chokepoint does It models a cloud environment as a graph in FalkorDB and gives an agent graph tools to: 1. Find every attack path from an entry point to a crown jewel. At depth 2 and 3 there is nothing. At depth 4, a forgotten service account (svc-legacy-etl) opens a route to prod-customer-db. 19 paths in total, each shown hop by hop. 2. Rank the edges worth cutting. The trap: one edge sits on 15 of 19 paths but costs 9 in disruption. Counting paths alone picks it. Chokepoint weighs paths cut against disruption instead. 3. Plan the fix: 2 changes cut 19 of 19 paths at a total disruption of 5. 4. Prove it before anything is touched. It copies the graph, applies the fix to the copy, re-runs the search, and shows 19 paths become 0. Then the scratch graph is deleted. 5. Propose the change with an owner team (found by traversal), a ticket and a config diff. 6. Wait for a human. Approving or rejecting is only ever a human click. Reject a change and the reason is written to the graph as a Decision node, so the next plan excludes that edge and finds an alternative. How FalkorDB is used (the graph is the product, not decoration) - One PIVOT relationship type with kind, cost and disruption properties - Variable-length Cypher traversal for path enumeration - UNWIND relationships(p) with count(DISTINCT p) for choke ranking - algo.SPpaths for weighted reachability, algo.pageRank and algo.betweenness for load-bearing assets - Graph copy and delete for the scratch simulation - Decision nodes linked to assets for memory that survives between sessions Design choices I care about - The model never writes Cypher. It chooses from a fixed catalogue of human-written, parameterised tools over MCP. - Every asset name is resolved against the graph first. An unknown name runs no traversal at all. - An answer checker rejects any model sentence that names an asset the graph did not return. - If there is no path, it says "no modelled path", never "safe". An empty result means the graph does not know, which is not the same as secure. - It runs on free tiers, and everything also works with no LLM at all through a deterministic router. Honest limits The planner is greedy, not provably optimal. All data is synthetic. A real deployment would need ingestion from real cloud configurations. Try it Live: trychokepoint.vercel.app Code: github.com/Im-2/chokepoint #cybersecurity