Charlie(@0xBroze):DeFi security has certainly been one of the most brutal parts of operating in onchain finance since inception, and the last 8 months have felt like a series of incessant blows…however, my hopeful view is that we are nearing the nadir, and the main threat / fear vector we face (agentic exploit, further ignited via Mythos concerns, etc) will soon actually become the primary protection layer for open, onchain financial protocols Essentially 2 theses here: 1. Battle-tested protocols that are immutable or with absolute minimal surface area for change will compound in value over the coming years. Strong risk management at the asset/parameter level is a given, but the compounding value comes via the contracts 2. We end up in a drone-warfare-state-equivalent but for onchain protocols where agents battle one another at the expense of tokens but no longer at the expense of user funds. These agent wars we witness create catalysts to double down on battle-tested protocols as well as ensure any remaining holes no longer linger for months. Can expand more on this in a future tweet Certainly other top teams working on DeFi security in different ways like @blockaid_, @phylaxsystems, @chainalysis, @HypernativeLabs, @HackenProof (multiple of whom we work with as well), but I think these two pieces above could be two of the keys over the next year+ to solving the constant lingering fear of fund loss due to random exploit Plan on our end at Felix is to double-down on #1 via strictly operating with battle-tested contracts and help move the needle forward on #2 as quickly as possible

2026.07.15 14:09

DeFi security has certainly been one of the most brutal parts of operating in onchain finance since inception, and the last 8 months have felt like a series of incessant blows…however, my hopeful view is that we are nearing the nadir, and the main threat / fear vector we face (agentic exploit, further ignited via Mythos concerns, etc) will soon actually become the primary protection layer for open, onchain financial protocols Essentially 2 theses here: 1. Battle-tested protocols that are immutable or with absolute minimal surface area for change will compound in value over the coming years. Strong risk management at the asset/parameter level is a given, but the compounding value comes via the contracts 2. We end up in a drone-warfare-state-equivalent but for onchain protocols where agents battle one another at the expense of tokens but no longer at the expense of user funds. These agent wars we witness create catalysts to double down on battle-tested protocols as well as ensure any remaining holes no longer linger for months. Can expand more on this in a future tweet Certainly other top teams working on DeFi security in different ways like @blockaid_, @phylaxsystems, @chainalysis, @HypernativeLabs, @HackenProof (multiple of whom we work with as well), but I think these two pieces above could be two of the keys over the next year+ to solving the constant lingering fear of fund loss due to random exploit Plan on our end at Felix is to double-down on #1# via strictly operating with battle-tested contracts and help move the needle forward on #2# as quickly as possible

Felix@felixprotocol

2026.07.15 14:05

Felix agent security Over the past 8 months, a series of exploits, multiple likely using frontier AI models to find frontend and smart contract-level vulnerabilities in DeFi apps, have altered the DeFi risk landscape gravely. This has led to a clear re-underwriting of acceptable DeFi risk and yield profiles. To combat this threat, Felix Labs has been putting in place what we are terming for now our agent security system. Concretely, this means: • Every code change is reviewed by the latest available frontier models before deployment, alongside our existing audit and human review process. • Ongoing model-driven review of the deployed Felix codebase and our OpSec processes, re-run as newer models become available. • Reviews operating alongside our existing monitoring systems covering both system risk and collateral risk. No security system is a guarantee. However, what this does accomplish is 1) meaningfully raise the cost for a malicious actor and 2) adds a detection layer that improves with every model generation. Risk and security continue to be our priority #1# at Felix. Offering the best yields to Hyperliquid lenders and the best capital efficiency to borrowers are both key value propositions but secondary if protocol solvency is not first in check. Feel free to reach out if interested in hearing more on our systems.

Forward to community