AI labs including Anthropic, OpenAI, and Google are proposing third-party auditors to verify AI safety practices, but security experts argue the labs should instead focus on basic network security controls like logging, permissions, and sandboxing to prevent AI models from accessing the internet and penetrating external systems during training.
ChainGPT discusses how automated triage systems for smart contract auditing create security vulnerabilities, as code flagged as low-risk bypasses human review. The platform offers an auditing service with formal verification to address this gap.
Respawn is a Rust-based content-addressed filesystem snapshot tool that enables versioning, undoing, and reverting directory states with cryptographic integrity guarantees. It supports atomic per-file reverts, tamper-evident auditing, and peer-to-peer replication over LAN without a central server, designed for safely rolling back changes made by autonomous processes.
Generate-Ilograph is a Claude Code skill that automatically creates interactive, auditable sequence diagrams from codebases using YAML syntax. The diagrams include code references, support nesting for complex flows, and can be updated as the codebase evolves, with output compatible with Ilograph web, desktop, and Confluence Cloud versions.
Trail of Bits operates one of three independent auditors for Signal's new Automatic Key Verification system, which uses key transparency to prevent servers from secretly substituting false public keys during encrypted chats. The auditor maintains a Merkle tree of the user-to-public-key map and cryptographically signs it to ensure all clients see the same consistent view, limiting any potential attack to one week before warnings appear.