PREVIEW. This is how it will look in an inbox; it was not sent to you. Each real email carries its own unsubscribe link and Keryn's postal address. KERYN HEADLINES Daily, Wednesday 19 August 2026 Retrospective Covers 18 August 2026 00:00 to 19 August 2026 00:00 UTC Web: https://keryn.dev/headlines/daily/2026-08-19/ TOP DEVELOPMENTS [1] Microsoft Copilot reveals secret input that allowed it to be hacked Ars Technica: Security · Incident · https://arstechnica.com/security/2026/08/microsoft-copilot-reveals-secret-input-that-allowed-it-to-be-hacked [2] npm v12 Lifecycle Script Limits: A Real Malicious Package Risk Reduction, or Just Moving Risk Around? Checkmarx · Incident · https://checkmarx.com/zero-post/npm-v12-lifecycle-script-limits-a-real-malicious-package-risk-reduction-or-just-moving-risk-around [3] 16 Typosquatted RubyGems Packages Steal Browser Credentials and Crypto Wallets The Hacker News · Research · https://thehackernews.com/2026/08/16-typosquatted-rubygems-packages-steal.html SAST [4] BullsEye: Directed Firmware Fuzzing arXiv · Research · https://arxiv.org/abs/2608.17729 [5] Benchmarking Automated Security Patch Backporting: How Far Are We? arXiv · Research · https://arxiv.org/abs/2608.17671 [6] SNIPTEST: Fuzzing Multi-Level Code Slices for Validating Vulnerabilities arXiv · Research · https://arxiv.org/abs/2608.17396 DAST [7] COMMITGUARD: Differential Slice Fuzzing for Commit-Induced Bug Detection arXiv · Research · https://arxiv.org/abs/2608.17401 AGENTIC SDLC [8] HarnessRisk: A Lifecycle-Oriented Benchmark for Agent Harness Safety arXiv · Research · https://arxiv.org/abs/2608.17597 [9] Microsoft Copilot Personal Flaws Could Let One Click Exfiltrate Data From Connected Apps The Hacker News · Analysis · https://thehackernews.com/2026/08/microsoft-copilot-personal-flaws-could.html REMEDIATION [10] Benchmarking Secure-and-Functional Remediation and How Snyk Agent Fix Lifts Frontier-Model Fix Rates by over 14% Snyk · Analysis · https://snyk.io/blog/snyk-agent-fix-remediation-benchmark MODELS [11] MemCatalyst: Amplifying Data Auditing on Vision-Language Models via Data Poisoning arXiv · Research · https://arxiv.org/abs/2608.17722 [12] Reflex-Guard: A Low-Latency Guardrail for LLM Prompt Safety Using Dense Semantic Embeddings arXiv · Research · https://arxiv.org/abs/2608.17556 [13] Fair ASR: Re-Evaluating Black-Box Jailbreaks under Shared Target-Call Budgets arXiv · Research · https://arxiv.org/abs/2608.17360 AI SECURITY [14] MobileWorldSafety: Benchmarking GUI Agent Safety Against Environmental Injection Attacks in Android Apps arXiv · Research · https://arxiv.org/abs/2608.17659 [15] When Agents Act on Web3: An Attack-Surface Survey of MCP, Skills, and Tool Calling arXiv · Research · https://arxiv.org/abs/2608.17275 [16] PACE: Policy-Attested Contract Execution for Safe AI Agents in Decentralized Finance arXiv · Research · https://arxiv.org/abs/2608.17220 ADVISORIES IN THIS WINDOW 56 npm advisories · 50 critical · 2 high · 3 moderate · 1 low 20 PyPI advisories · 5 critical · 6 high · 8 moderate · 1 low 11 Packagist advisories · 2 critical · 5 high · 4 moderate 8 Maven advisories · 1 critical · 4 high · 2 moderate · 1 low 3 Go advisories · 1 high · 2 moderate 1 crates.io advisory · 1 moderate 1 Hex advisory · 1 moderate 1 RubyGems advisory · 1 critical 1 Multi-ecosystem advisory · 1 high All advisories: https://keryn.dev/headlines/advisories/ Titles, links and short excerpts only; every item links to its source. Preview: the unsubscribe link and Keryn's postal address are added to each email when it is sent.