Panther Audits — Security ResearchMove-first smart contract security field notes, findings, formal verification, and research labs.https://pantheraudits.com/blog/2026-08-19T00:00:00ZPantherhttps://pantheraudits.com/Beyond the Bet: Auditing a Prediction Market on Sui Movehttps://pantheraudits.com/blog/when-odds-become-state-auditing-a-prediction-market-on-sui-move.html2026-08-19T00:00:00Z2026-08-19T00:00:00ZField notes from a private one-month review of a Sui Move prediction market and its signed oracle integration.Between Two Chains: How I Audit Cross-Chain DeFi Protocolshttps://pantheraudits.com/blog/how-to-audit-cross-chain-defi-protocols.html2026-07-16T00:00:00Z2026-08-05T00:00:00ZLessons from a private EVM audit on asynchronous bridges, mirrored accounting, stale snapshots, recovery paths, and cross-chain invariants.When Collateral Is a Live Market: Auditing Lending Protocols on Bittensorhttps://pantheraudits.com/blog/how-to-audit-lending-protocols-on-bittensor.html2026-07-16T00:00:00Z2026-07-16T00:00:00ZLessons from a private Bittensor lending audit covering unit-safe accounting, Dynamic TAO pricing, staking precompiles, liquidation liveness, and runtime-aware testing.Private Bug Bounty Triage: Validating and Reproducing Web3 Findingshttps://pantheraudits.com/blog/private-bug-bounty-triage.html2026-06-05T00:00:00Z2026-06-05T00:00:00ZHow I validate private Web3 findings during bounty triage, then write deterministic PoCs, impact evidence, and reviewer-ready artifacts.How to Audit a Perp Protocol on Sui Movehttps://pantheraudits.com/blog/how-to-audit-a-perp-protocol-on-sui-move.html2026-04-15T00:00:00Z2026-04-15T00:00:00ZAuditing perps on Sui requires a different checklist from EVM or Aptos. The object model, hot-potato request/response pattern, Programmable Transaction Blocks, and shared collateral vaults all reshape the attack surface — the Sui companion to the…The Upgrade Trap — How Versioned Objects Can Brick Protocols on Suihttps://pantheraudits.com/blog/the-upgrade-trap-how-sui-move-versioned-objects-can-brick-protocols.html2026-04-02T00:00:00Z2026-04-02T00:00:00ZSui's Versioned wrapper lets protocols migrate shared objects across upgrades. But the same pattern creates an upgrade trap — if your immutable package depends on a versioned protocol, one migrate() call bricks you forever.Move VM Runtime — How Your Move Code Actually Runshttps://pantheraudits.com/blog/move-vm-runtime-explained.html2026-04-01T00:00:00Z2026-04-01T00:00:00ZHow the Move VM runtime actually executes your code — from bytecode loading to interpreter dispatch and native function calls.Sui Execution Layer — A Security Researcher's Deep Divehttps://pantheraudits.com/blog/sui-execution-layer-deep-dive.html2026-03-30T00:00:00Z2026-03-30T00:00:00ZA deep dive into Sui's execution layer — how transactions flow from submission to state changes, and what security researchers should look for.Sui Bella Ciao — Inside the New Move VMhttps://pantheraudits.com/blog/sui-bella-ciao-new-move-vm.html2026-03-28T00:00:00Z2026-03-28T00:00:00ZSui rewrote their entire Move VM. Here's how the new execution layer works, what changed from the old one, and where the interesting attack surface lives for security researchers.Sui's Cut Package — How Sui Freezes Its Execution Layerhttps://pantheraudits.com/blog/sui-execution-cut-package-explained.html2026-03-28T00:00:00Z2026-03-28T00:00:00ZHow Sui uses the cut package mechanism to freeze its execution layer — snapshotting framework code at specific protocol versions for deterministic replay.ML04: Membership Inference Attack — 0xTheBlackPantherhttps://pantheraudits.com/blog/ai-security-ML04-membership-inference-attack.html2026-03-26T00:00:00Z2026-03-26T00:00:00ZYou don't need to extract the training data. You just need to ask the right questions — and the model will tell you whose data it learned from.ML05: Model Theft — 0xTheBlackPantherhttps://pantheraudits.com/blog/ai-security-ML05-model-theft.html2026-03-26T00:00:00Z2026-03-26T00:00:00ZYou don't need to break into anyone's servers. If a model is queryable, it is reconstructable.ML06: ML Supply Chain Attacks — 0xTheBlackPantherhttps://pantheraudits.com/blog/ai-security-ML06-ML-supply-chain-attacks.html2026-03-26T00:00:00Z2026-03-26T00:00:00ZYou audited your model. You reviewed your code. But did you audit every library your training script imported?ML07: Transfer Learning Attack — 0xTheBlackPantherhttps://pantheraudits.com/blog/ai-security-ML07-transfer-learning-attack.html2026-03-26T00:00:00Z2026-03-26T00:00:00ZThe model passed every benchmark. It scored well on your validation set. But it was trained to fail in exactly the one scenario the attacker controls.ML08: Model Skewing — 0xTheBlackPantherhttps://pantheraudits.com/blog/ai-security-ML08-model-skewing.html2026-03-26T00:00:00Z2026-03-26T00:00:00ZThe model wasn't broken on day one. It was steered — slowly, carefully, through fake feedback — until its decisions aligned with what the attacker needed.ML09: Output Integrity Attack — 0xTheBlackPantherhttps://pantheraudits.com/blog/ai-security-ML09-output-integrity-attack.html2026-03-26T00:00:00Z2026-03-26T00:00:00ZEvery other attack tampers with the model. This one doesn't touch it at all. It intercepts the correct answer on the way out — and replaces it with a lie.ML10: Model Poisoning — 0xTheBlackPantherhttps://pantheraudits.com/blog/ai-security-ML10-model-poisoning.html2026-03-26T00:00:00Z2026-03-26T00:00:00ZEvery other attack manipulates what the model learns or outputs. This one goes straight to the parameters and rewrites them directly.ML02: Data Poisoning Attack — 0xTheBlackPantherhttps://pantheraudits.com/blog/ai-security-ML02-data-poisoning-attack.html2026-03-25T00:00:00Z2026-03-25T00:00:00ZThe attack that happens before the model even wakes up — how data poisoning corrupts AI at the weights level.ML03: Model Inversion Attack — 0xTheBlackPantherhttps://pantheraudits.com/blog/ai-security-ML03-model-inversion-attack.html2026-03-25T00:00:00Z2026-03-25T00:00:00ZYou don't need to steal the training data. You just need an API and patience — and the model will hand it back to you.ML01: Input Manipulation Attack — 0xTheBlackPantherhttps://pantheraudits.com/blog/ai-security-ML01-input-manipulation-attack.html2026-03-24T00:00:00Z2026-03-24T00:00:00ZWhat happens when the model's reality and human reality are different things — and an attacker exploits that gap.How to Audit Aave V3 Fork Deployment Scripts — The Unsexy Attack Surfacehttps://pantheraudits.com/blog/how-to-audit-aave-v3-fork-deployment-scripts.html2026-03-22T00:00:00Z2026-03-22T00:00:00ZLearnings from auditing Aave V3 fork deployment scripts — the bug classes, the patterns, and the real findings from production audits.How Lending Protocols Work On-Chain — And How to Audit Them on Sui Movehttps://pantheraudits.com/blog/how-to-audit-a-lending-protocol-on-sui-move.html2026-03-16T00:00:00Z2026-03-16T00:00:00ZShare-ratio manipulation, eMode edge cases, oracle staleness, and the liquidation paths that actually break — a field checklist from a live Sui lending audit.Formal Verification of Aptos Move Contracts Using the Move Proverhttps://pantheraudits.com/blog/formal-verification-of-aptos-move-contracts-using-move-prover.html2026-03-15T00:00:00Z2026-03-15T00:00:00ZSetting up the Move Prover on a real Aptos codebase: writing spec blocks, proving aborts_if conditions, and the invariants that catch bugs manual review misses.Formal Verification of Sui Move Lending Protocol — Part 2: eMode Borrow Restrictions & Finding a Bug with Certorahttps://pantheraudits.com/blog/formal-verification-of-current-sui-using-certora-part-2.html2026-03-14T00:00:00Z2026-03-14T00:00:00ZUsing Certora's Sui Prover to encode borrow restrictions as parametric rules — and what the prover surfaced that manual review walked past.Formal Verification of Sui Move Lending Protocol Using Certora Proverhttps://pantheraudits.com/blog/formal-verification-of-current-sui-using-certora-sui-prover.html2026-03-13T00:00:00Z2026-03-13T00:00:00ZHow I mathematically proved that an eMode group assignment is immutable after obligation creation in a Sui DeFi lending protocol — using Certora's Sui Prover, parametric rules, and field accessors.ZK Journey Week 5 — Isomorphisms, Why Cubics, and Counting Pointshttps://pantheraudits.com/blog/zk-journey-week-05.html2026-03-12T00:00:00Z2026-03-12T00:00:00ZAs a Move security researcher, I need to understand ZK proofs deeply — not just what they do, but how they work under the hood. Week 5 of my ZK learning journey at the Rare Skills ZK Bootcamp.ZK Journey Week 4 — Elliptic Curves: Drawing Lines, Building Groupshttps://pantheraudits.com/blog/zk-journey-week-04.html2026-03-09T00:00:00Z2026-03-09T00:00:00ZWeek 4 of my ZK learning journey at the Rare Skills ZK Bootcamp.How to Audit a Gold-Backed Staking Protocol on Solanahttps://pantheraudits.com/blog/how-to-audit-a-gold-backed-staking-protocol-on-solana.html2026-03-08T00:00:00Z2026-03-08T00:00:00ZRWA-backed staking on Solana: mint/redeem trust boundaries, reward accounting, oracle dependencies, and the Anchor account-validation traps from a real engagement.How to Audit a Diamond-Pattern Stablecoin Protocol — Lessons from a Real Engagementhttps://pantheraudits.com/blog/how-to-audit-a-stablecoin-protocol.html2026-03-06T00:00:00Z2026-03-06T00:00:00ZPeg mechanics, collateral accounting, mint/burn access paths, and diamond-pattern upgrade risk — what I check first in every stablecoin review.I Built an Open-Source Move Auditor Skill for Claude Codehttps://pantheraudits.com/blog/the-move-auditor.html2026-03-05T00:00:00Z2026-03-05T00:00:00ZIntroducing move-auditor — an open-source Claude skill that runs a structured security-audit pipeline over Move codebases, built from my real engagement workflow.ZK Journey Week 3 — Why Prime? From Group Theory to Real Cryptographyhttps://pantheraudits.com/blog/zk-journey-week-03.html2026-03-04T00:00:00Z2026-03-04T00:00:00ZWeek 3 of my ZK learning journey at the Rare Skills ZK Bootcamp.I Open-Sourced My AI Audit Workflow — Prompts, Pipelines, and Everything I Learnedhttps://pantheraudits.com/blog/i-open-source-my-audit-prompts-and-workflow.html2026-03-03T00:00:00Z2026-03-03T00:00:00ZThe full prompt pipeline I use to assist real audits — recon, hypothesis generation, verification gates — published with notes on what worked and what didn't.ZK Journey Week 2 — Abstract Algebra: From Sets to Fieldshttps://pantheraudits.com/blog/zk-journey-week-02.html2026-03-02T00:00:00Z2026-03-02T00:00:00ZWeek 2 of my ZK learning journey at the Rare Skills ZK Bootcamp.How Perpetual Futures Work On-Chain — And How to Audit Themhttps://pantheraudits.com/blog/how-to-audit-a-perp-protocol-on-aptos-move.html2026-03-01T00:00:00Z2026-03-01T00:00:00ZLearnings from auditing a fully on-chain perp DEX in Move on Aptos — orderbook mechanics, oracle integration, liquidation engine, funding rate, and the bugs that actually matter.Field Order Is Sort Order — How Struct Layout Silently Breaks Ordered Maps in Movehttps://pantheraudits.com/blog/silent-bug-in-big-ordered-map-struct-move.html2026-02-23T00:00:00Z2026-02-23T00:00:00ZMove compares structs lexicographically by field declaration order. If you use a struct as a key in an ordered map, the first field dominates sorting — not the field you think matters. A real bug from Decibel Exchange's perpetual futures DEX on…The Delete-Then-Validate Bug — How return vs abort Silently Corrupts State in Movehttps://pantheraudits.com/blog/the-silent-killer-in-move-how-return-after-state-mutation-permanently-corrupts-on-chain-data.html2026-02-22T00:00:00Z2026-02-22T00:00:00ZA real bug from Aptos core's trading engine that permanently deleted orders on an "error" path. The root cause? A return where there should have been an abort. Applies to both Aptos and Sui Move.ZK Journey Week 1 — Why a Move Security Researcher Is Learning Zero Knowledge Proofshttps://pantheraudits.com/blog/zk-journey-week-01.html2026-02-20T00:00:00Z2026-02-20T00:00:00ZAs a Move security researcher, I need to understand ZK proofs deeply — not just what they do, but how they work under the hood. Week 1 of my ZK learning journey at the Rare Skills ZK Bootcamp.The Wrong Address: The Move Bug That Would Have Bricked Aave on Aptoshttps://pantheraudits.com/blog/the-move-bug-that-would-have-bricked-aave-on-aptos.html2026-02-14T00:00:00Z2026-08-05T00:00:00ZPanther placed 3rd in Aave Aptos V3 and received 24,293.09 GHO total contest payout for 1 confirmed High and 1 confirmed Medium. This case study explains the protocol-bricking High.From the Trenches — Lessons from 6x Top 3 Finishes in Competitive Auditshttps://pantheraudits.com/blog/how-to-win-a-public-contest.html2026-02-12T00:00:00Z2026-02-12T00:00:00ZZero fluff, pure trenches-tested insights from climbing to the top 50s on Cantina. This is the alpha that actually moved the needle — not theory, not vibes, just what works when you're deep in the code at 3 AM and the leaderboard is watching.The Ghost Owner — CVE-2024-45304https://pantheraudits.com/blog/the-ghost-owner.html2026-02-11T00:00:00Z2026-08-05T00:00:00ZCVE-2024-45304 in OpenZeppelin Cairo OwnableTwoStep left pending ownership live after renunciation, allowing a pending owner to reclaim control. This case study covers the root cause, impact, and fix.