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, Thursday 3 September 2026 Retrospective Covers 2 September 2026 00:00 to 3 September 2026 00:00 UTC Web: https://keryn.dev/headlines/daily/2026-09-03/ TOP DEVELOPMENTS [1] BGP hijack infecting networks caused by a comedy of errors that’s not funny at all Ars Technica: Security · Incident · https://arstechnica.com/security/2026/09/well-executed-bgp-attack-uses-hijacked-ips-to-infect-real-networks [2] CVE-2026-49869: Kestra OSS OS Command Injection Vulnerability added to CISA's Known Exploited Vulnerabilities catalog CISA · Advisory · https://cisa.gov/known-exploited-vulnerabilities-catalog?search_api_fulltext=CVE-2026-49869 [3] CVE-2026-9586: Sangoma Switchvox SQL Injection Vulnerability added to CISA's Known Exploited Vulnerabilities catalog CISA · Advisory · https://cisa.gov/known-exploited-vulnerabilities-catalog?search_api_fulltext=CVE-2026-9586 SAST [4] Automated Vulnerability Injection in Smart Contracts Using Large Language Models arXiv · Research · https://arxiv.org/abs/2609.02624 SCA [5] AgOSS: A Dataset and Multi-Layer Characterization of Open-Source Agricultural Software arXiv · Research · https://arxiv.org/abs/2609.02591 [6] Stored Is Not Supported: Typed Provenance and Assertion Guardrails for Persistent AI Agents arXiv · Research · https://arxiv.org/abs/2609.02127 [7] Get the Spectra Assure Community Plugin for JFrog Artifactory ReversingLabs · Analysis · https://reversinglabs.com/blog/spectra-assure-community-plugin-for-jfrog-artifactory DAST [8] Burp Suite vs OWASP ZAP: Which Fits Your Testing Workflow? Aikido Security · Analysis · https://aikido.dev/blog/burp-suite-vs-owasp-zap AGENTIC SDLC [9] Malicious .git Configs Can Make Claude, Codex, Cursor, and Other AI Agents Run Attacker Code The Hacker News · Analysis · https://thehackernews.com/2026/09/malicious-git-configs-can-make-claude.html REMEDIATION [10] Attackers Exploit Two SonicWall SMA 1000 Zero-Days That May Form an Attack Chain The Hacker News · Analysis · https://thehackernews.com/2026/09/attackers-exploit-two-sonicwall-sma.html TRUST [11] Open Problems in AI Risk Modeling: Insights from a Workshop on the Technical Foundations of AI Risk Modeling arXiv · Research · https://arxiv.org/abs/2609.03178 [12] PrimSynth: An Agentic Approach to Discover, Validate, and Synthesize Exploit Primitives for Linux Kernel Vulnerabilities arXiv · Research · https://arxiv.org/abs/2609.02647 [13] The Defense Rests: A Courtroom-Inspired Multi-Agent LLM Framework for Automated Safety Case Generation arXiv · Research · https://arxiv.org/abs/2611.21618 MODELS [14] Beyond Small Patches: Black-Box Detection and Purification of Diverse Backdoor Triggers arXiv · Research · https://arxiv.org/abs/2609.03139 [15] SafeEvolve: Harness-Policy Co-Evolution from Agent Experience for Safety Alignment arXiv · Research · https://arxiv.org/abs/2609.02786 [16] SEAL: Reinforcing Global Safety in Mixture-of-Experts through Shared Expert ALignment arXiv · Research · https://arxiv.org/abs/2609.02293 AI SECURITY [17] One Question, Four Voices: How Advice for Alzheimer's Caregiving Differs Between Caregivers, Clinicians, and Large Language Models arXiv · Research · https://arxiv.org/abs/2611.18604 [18] ACLE-MCP: Attested Capability Leases for Execution-Time Trust in Remote LLM Tool Use arXiv · Research · https://arxiv.org/abs/2609.02690 [19] Implicit Manipulation for Skill Selection in LLM Agents with Semantic Matching arXiv · Research · https://arxiv.org/abs/2609.02035 ADVISORIES IN THIS WINDOW 9652 npm advisories · 9632 critical · 10 high · 10 moderate 18 PyPI advisories · 2 critical · 8 high · 8 moderate 12 Packagist advisories · 4 high · 8 moderate 11 Go advisories · 3 critical · 7 high · 1 moderate 1 Maven advisory · 1 high 1 crates.io advisory · 1 moderate 1 RubyGems advisory · 1 moderate 1 NuGet advisory · 1 low 1 Multi-ecosystem advisory · 1 moderate 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.