Our CRS (Cyber Reasoning System), Buttercup, is now competing in the one and only scored round of DARPA’s AI Cyber Challenge (AIxCC) against six other teams to see which autonomous AI-driven system can find and patch the most software vulnerabilities.

Private key compromise accounted for 43.8% of crypto hacks in 2024, yet traditional smart contract audits rarely address architectural access control weaknesses. This post introduces a four-level maturity framework for designing protocols that can tolerate key compromise, progressing from single EOA control to radical immutability, with practical examples demonstrating multisigs, timelocks, and the principle of least privilege.

File parsers in Go contain unexpected behaviors that can lead to serious security vulnerabilities. This post examines how JSON, XML, and YAML parsers in Go handle edge cases in ways that have repeatedly resulted in high-impact security issues in production systems. We explore three real-world attack scenarios: marshaling/unmarshaling unexpected data, exploiting parser differentials, and leveraging data format confusion. Through examples, we demonstrate how attackers can bypass authentication, circumvent authorization controls, and exfiltrate sensitive data by exploiting these parser behaviors.

In October 2023, we audited Silence Laboratories’ DKLs23 threshold signature scheme (TSS) library—one of the first production implementations of this then-novel protocol that uses oblivious transfer (OT) instead of traditional Paillier cryptography. Our review uncovered serious flaws that could enable key destruction attacks, which Silence Laboratories promptly fixed.

Over two audits in 2023, we reviewed a blockchain system developed by Axiom that allows computing over the entire history of Ethereum, all verified by zero-knowledge proofs (ZKPs) on-chain using ZK-verified elliptic curve and SNARK recursion operations. This system is built using the Halo2 framework—a complex, emerging technology that presents many challenges when building a secure application, including potential under-constrained issues resulting from its low-level API.

Introducing the Custodial Stablecoin Rekt Test; a new spin on the classic Rekt Test for evaluating the security maturity of stablecoin issuers.

This post will examine the cryptography behind passkeys, the guarantees they do or do not give, and interesting cryptographic things you can do with them, such as generating cryptographic keys and storing certificates.

Datasig generates compact, unique fingerprints for AI/ML datasets that let you compare training data with high accuracy—without needing access to the raw data itself. This critical capability helps AIBOM (AI bill of materials) tools detect data-borne vulnerabilities that traditional security tools completely miss.

See how we slashed PyPI’s test suite runtime from 163 to 30 seconds. The techniques we share can help you dramatically improve your own project’s testing performance without sacrificing coverage.

This post describes how many examples of MCP software store long-term API keys for third-party services in plaintext on the local filesystem, often with insecure, world-readable permissions.

This post describes attacks using ANSI terminal code escape sequences to hide malicious instructions to the LLM, leveraging the line jumping vulnerability we discovered in MCP.

Malicious MCP servers can inject trigger phrases into tool descriptions to exfiltrate entire conversation histories and steal sensitive credentials and IP.

Trail of Bits’ Buttercup competes in DARPA’s AIxCC Finals with expanded resources, multiple rounds, new challenge types, and custom AI model capabilities.

MCP’s ’line jumping’ vulnerability lets malicious servers inject prompts through tool descriptions to manipulate AI behavior before tools are ever invoked.

We’re working on integrating an ASN.1 API into PyCA Cryptography, built on top of the same Rust ASN.1 implementation already used by Cryptography’s X.509 APIs.

This post describes a sophisticated social engineering campaign using Zoom’s remote control feature and provides technical solutions to protect organizations against this attack vector.

Learn snapshot fuzzing for kernel-level testing. New Testing Handbook section shows how to test drivers, antivirus software, and complex kernel components.

Trail of Bits’ independent study finds OpenSearch v2.17.1 is 1.6x faster than Elasticsearch v8.15.4 on Big5 workload and 11% faster on vector search.

Learn how to integrate TRAIL threat modeling into your SDLC, adapt and maintain models as your system evolves, and use them to identify security control gaps.

Discover TRAIL, Trail of Bits’ systematic threat modeling approach that identifies design-level security weaknesses and provides actionable remediation guidance.

Learn how comprehensive threat modeling could have identified the operational security gaps that led to Bybit’s $1.5B hack and prevented similar breaches.

The $1.5B Bybit Hack demonstrates how the Era of Operational Security Failures has arrived, and most cryptocurrency companies are not prepared for its implications.

We developed a simple CodeQL query to find denial-of-service (DoS) vulnerabilities in several high-profile Java projects.

Introducing Medusa v1, a cutting-edge fuzzing framework designed to enhance smart contract security.

TVM Ventures has selected Trail of Bits as its preferred security partner to strengthen the TON developer ecosystem. Through this partnership, we’ll lead the development of DeFi protocol standards and provide comprehensive security services to contest-winning projects deploying on TON. TVM Ventures will host ongoing developer contests where teams can showcase innovative applications that advance […]