Engineering briefs Claude Infrastructure Tampering: Emergent Reward Hacking A Claude Opus 4.6 manager agent modified scoring infrastructure when workers repeatedly failed quality gates. This case study examines the incident and the controls needed to protect the delivery pipeline. Read more →
Engineering briefs PHP’s Critical Flaws Reinforce Why Assume Breach Must Be the Default PHP’s critical flaws show why architects must plan for successful intrusion, limit attacker options, and preserve controls beyond the compromised service. Read more →
Engineering briefs How Your Production Code Enables the AI Cyber-Apocalypse Readable code and instructions help developers build and repair systems. AI helps attackers understand them too. Once inside, they can use that knowledge to steal data, divert payments, and disrupt services. Read more →
Engineering briefs Building software with an AI delivery pipeline How specialised AI agents build software through a structured pipeline—from requirements and tests to independent review, acceptance and maintenance. Read more →
Engineering briefs When software becomes a custodian. A Sarolta engineering case study: governed AI-native software delivery with durable task orchestration and production evidence. Read more →
Engineering briefs Your Code Is Disposable. Get Used to It. AI has made software easier to regenerate. The durable asset is the specification: the decisions, constraints, and evidence that make a system worth rebuilding. Read more →