A look at Anthropic’s Claude Code Security: its technical approach, false‑positive mitigations, the GitHub Action, comparisons with competing tools, and why $15B briefly vanished from cybersecurity stocks.
Socket reports an active campaign using 19 malicious npm packages. It targets AI development environments such as Claude, Cursor, and VS Code, stealing SSH keys, npm tokens, and API keys, and then propagates via a worm.
Techniques and defenses from the MINJA, InjecMEM, and ToxicSkills campaigns that poison AI agents’ memory files, and the fact that GPT-5.3-Codex achieved a 72% exploit success rate on EVMbench released by OpenAI and Paradigm. This article organizes how AI becomes both a target of attacks and a weapon for attackers.
Two arguments: a renewed look at Web Components asking ‘Do we really need React?’ and a push to ‘turn Dependabot off and switch to Go’s vulnerability checker.’ Both revisit long‑standing defaults with technical reasoning.
An intrusion campaign that auto-scanned FortiGate in 106 countries using DeepSeek and Claude; Starkiller, a reverse-proxy PhaaS that nullifies MFA; Anthropic's Claude Code Security finding 500+ vulnerabilities in production OSS; and PayPal exposing SSNs for six months due to a coding mistake.
Stripe Minions, Amazon Kiro, Claude Code compaction, and a Replit DB deletion. We synthesize multiple cases to extract the design principles required to operate AI coding agents in production, and organize them alongside CodeRabbit's 470‑repo statistics plus efforts from Google and GitHub.
In the same week, CISA's KEV catalog gained a Chromium CSS engine UAF, a Roundcube RCE that hid for over a decade, a BeyondTrust RCE abused by ransomware, and a Dagu RCE due to no default authentication. All four require immediate patching.
Andrej Karpathy coined "Claws" as an upper layer for AI agents, and June Kim answered the same question from a different angle with the Cord framework implemented with MCP and SQLite. This piece organizes the shift from single-shot agents to autonomous coordination systems from both conceptual and implementation perspectives.
Kiro autonomously deleted production, causing 13 hours of AWS downtime; Claude Code’s auto-compaction irreversibly erases context; sub-agents silently burn through usage. Three incident reports from the same week.
Stripe’s Minions agent generates 1,300+ PRs per week with zero human effort. Implementation details of the four components: Devbox, Blueprints, Toolshed, and a fork of goose.
A week that shook trust in both people and software: three former Google engineers were indicted over alleged transfers of sensitive information to Iran, while NetEase's MuMu Player was found running 17 reconnaissance commands on macOS every 30 minutes.
Two February 2026 papers on reducing inference cost: Together AI’s Consistency DLM (up to 14.5× faster) and MIT/Harvard’s Attention Matching KV compaction (50× compaction in seconds).
