FrontRunner🔥🛡💯

People who spent $700 on a mac mini to run an openclaw agent watching claude launch all the features natively for $20

@olanetsoft Happy birthday champ🏆

Happy birthday, Idris! Grateful for every lesson, every friendship, every win. The story's just getting started.

realizing AI is your generation's gold rush event, and 99% of your competition is still debating whether the gold even exists. (you're about to get filthy rich)

@MichaelAsiedu_ God when

Anytime my wife is ushering at church and our eyes meet, I make an ugly face, and she immediately breaks character. 😂 She is so sweet, and I love her so much.

✨Introducing evmresearch✨✨ A knowledge graph of nearly everything I've learned about the EVM in the past six years The graph structure emulates the brain, exponentiating research speeds for both humans and agents evmresearch.io

🚨BREAK!NG: Someone compiled every AI prompt you'll ever need in one place. You can find prompts for: - Acting as a Linux Terminal, Python Interpreter, SQL Console - Creative Writing, Screenwriting, Debate, Journalism - Career Coaching, Interview Prep, Resume Writing, Mentorship - Hundreds of community-tested prompts for any AI model 100% Open Source and available for everyone👇🏼

Announcing the Solidity Testing Handbook ✨ Fully free, one-stop resource for Solidity developers and security researchers. Resources are currently scattered across blogs, docs, and forums. I found it difficult to keep track of everything in one place. This handbook aggregates all testing patterns from basic unit tests to advanced mutation tests into a single, well-organized guide for quick reference. It’s built from my own learnings and best practices observed in popular codebases. soliditytestingbook.com

99% of developers use Codex wrong. They get garbage code and endless rewrites. Here’s how to turn it into a junior engineer who actually thinks 1. Stop prompting. Start building a task pipeline. Don’t throw one big prompt at Codex and hope it “figures it out.” Break work into discrete tickets: - Refactor this module. - Extract validation logic. - Improve error handling. - Write regression tests. Treat Codex like a junior engineer working off a structured backlog. Why this works: Large prompts create cognitive diffusion but smaller scoped tasks reduce ambiguity. Ambiguity is where the models hallucinate. When you think in tickets instead of prompts, output quality jumps immediately. 2. Create a permanent “brain” file for your project If you aren’t using a persistent context file (e.g., AGENTS.md, PROJECT_RULES.md, etc.), you’re resetting the model every session. Create one file that defines: - Architecture patterns - Naming conventions - Folder structure rules - Error-handling standards - State management decisions - Performance constraints Then reference it every time. Now Codex doesn’t guess your standards. It enforces them. This is how you move from “AI-generated code” to “AI operating inside your system.” 3. Force it to compete with itself (Best-of-N) Most people generate one solution and start editing. That’s super lazy. Instead, ask for: - 3 different architectural approaches - 3 different refactor patterns - 3 performance optimization strategies Then choose or merge the strongest parts. AI models are stochastic. The first output is rarely the best. If you force diversity, you increase surface area for better structure. 4. Always request a plan before requesting code. This is one of the biggest hidden upgrades. Don’t start with: “Write the feature.” Start with: “Outline the implementation plan. Which files change? What new modules are needed? What edge cases must be handled?” Then validate the plan. Then generate code. This does two things: - Reduces hallucinated file edits - Forces structural reasoning before token generation 5. Use it to map codebases, not just write them Don’t underuse Codex’s analytical capability. Instead of asking it to “add a feature,” try: - Trace the data flow from API layer to UI - Identify circular dependencies - List performance bottlenecks - Explain how authentication flows across services This turns Codex into an architecture explainer. Understanding the system before editing it reduces rewrites dramatically. 6. Parallelize tasks If your workflow allows it, split a large job into independent subtasks: - One instance writes tests - One instance refactors logic - One instance improves documentation Then merge manually. This mirrors how actual teams work. Sequential prompting wastes compute and time. Parallelization multiplies output velocity. 7. Fix your environment before blaming the model Many Codex failures are environment failures: - Broken dependencies - Unclear scripts - No linting - No test harness When the environment is structured, Codex becomes dramatically more reliable. Garbage context → garbage output. 8. Make it use real tools Instead of describing your repo state in text, have it: - Analyze git diff - Review test outputs - Interpret actual error logs - Run builds and linters Grounding in real artifacts beats abstract explanation every time. Models perform better when reasoning over concrete outputs instead of narrative descriptions.

A very interesting guide to AI Agents. services.google.com/fh/files/misc/…

@0x15_eth @code4rena @monad Congrats man💯🔥

Finally! After so many months result is out This was my first major win in blockchain/DLT last year and my first 5 fig contest win🔥 on @code4rena Happy to help secure @monad

@immunefi Bruh🔥

Security researcher ily2 has just earned a staggering $3,000,000 from submitting a critical smart contract bug via Immunefi. That's the largest single payout in web3 security in recent memory. In total, he's submitted 3 reports. All 3 were paid. 100% accuracy. His leaderboard update is coming soon, but you can pledge IMU to him now and earn when he finds the next one: immunefi.com/pledge/ily2

Can anyone recommend a good tool that lets you see all the state variables of a contract at a glance and all the functions that update those states. I would also appreciate if it shows external calls in one glance.

For those who may not qualify, I am on the path at the moment , so far the following has helped me: 1. @andrewbrown's GenAI Essentials Course : exampro.co/exp-genai-01 2. AI Engineering by @chipro 3. Standford's CME 295 by the Amidi Twins cme295.stanford.edu/syllabus/

@olanetsoft This is indeed very timely.

Over a year ago, everyone on my timeline was a Web 3 developer. Now, they're all AI engineers. In six months to a year, they'll be something else. I've stopped chasing trends. Here's what I do instead.

@0x15_eth Congrats man💯

Recent win...was really hoping this was gonna be a critical cos there was a lot of back and forth with the protocol team but after several e2e tests and constraints, i couldn't demonstrate a critical impact so i settled for demonstrating a medium instead Not bad cos it was a still a cool 5 fig payout...Still on the journey to find my first crit Lesson...test RIGOUROUSLY for the impact!

@raopreetam_ @QuillAudits DMs not open

We’re looking for 4-5 Security Audit Interns at @QuillAudits academy who don't just "read" code, but break it. The Stack: Solidity, Rust, Move. Nice to have: Hands-on experience with Testing & Fuzzing (Foundry, Echidna, Medusa). This is an unpaid 3-month internship designed as a high-octane trial. Perform well, and you’ll be fast-tracked into a Full-Time Auditor role. If you think like an attacker and build like a defender, let's talk

I'm Boris and I created Claude Code. I wanted to quickly share a few tips for using Claude Code, sourced directly from the Claude Code team. The way the team uses Claude is different than how I use it. Remember: there is no one right way to use Claude Code -- everyones' setup is different. You should experiment to see what works for you!

After 5 months full-time, first Web3sec W in the bag🤩🎉 Found 2H & 1L for just a few hours work on last day of audit, Glad for the result. Thank you @malik672_ for showing me Web3. Thank you @GuildAcademy_ for sharpening me. The best is yet to come 🔥😤 @0xSimao

Building a secure protocol is not that complicated… you just need to: - Incorporate security from the design - Discuss and address every edge case - Build a strong test case - Get the code audited by strong auditors (preferably multiple times) - Formally verify your code - Develop an extra safe frontend - Build a comprehensive and detailed emergency plan - Carefully build a deploy script (and ideally get it audited too) - Set up and manage a bug bounty program (including for the frontend) - Closely monitor the contracts immediately after launch to ensure the configs are correct and the deployment hasn’t been attacked - Perform a staged rollout, progressively increasing the limits - Monitor the chain for suspicious activity - Monitor every dependency that could break your protocol - Maintain an external safe failure mode (e.g. pausing guardians) - Properly manage governance - Properly manage signer keys And some other things