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, Friday 26 June 2026 · Retrospective · Covers 25 June 2026 00:00 to 26 June 2026 00:00 UTC

Read on the web

Top developments

Interesting Paper Exploring Prompt Injection

Schneier on Security · Research · AI Security

CVE-2026-12569: PTC Windchill and FlexPLM Improper Input Validation Vulnerability added to CISA's Known Exploited Vulnerabilities catalog

CISA · Advisory

New macOS malware embeds fake errors to confuse AI analysis tools

BleepingComputer · Analysis · AI Security

SCA

npm adds preventive account protection for high-impact accounts

GitHub · Analysis · SCA

Chai: Agentic Discovery of Cryptographic Misuse Vulnerabilities

arXiv · Research · SCA

Operation Navy Ghost: How Attackers Planted a Telegram-Powered Backdoor Across Fake pyrogram Packages on PyPI

Checkmarx · Analysis · SCA

DAST

Protocol Prying: Systematic Vulnerability Research in the Apple AirDrop and Android Quick Share Proximity Transfer Protocols

arXiv · Research · DAST

Agentic SDLC

How Much Static Structure Do Code Agents Need? A Study of Deterministic Anchoring

arXiv · Research · Agentic SDLC

A Deterministic Control Plane for LLM Coding Agents

arXiv · Research · Agentic SDLC

Knowledge-Based Pull Requests: A Trusted Workflow for Agent-Mediated Knowledge Collaboration

arXiv · Research · Agentic SDLC, AI Security, Trust

Remediation

When Passing Tests Hides Vulnerabilities: An Empirical Study of Silent Failures in Agentic Systems

arXiv · Research · Remediation

Trust

The CRA Readiness Reality: What Changed (and What Didn’t) Between 2025 and 2026?

OpenSSF · Standard / policy · Trust

Autoformalization of Agent Instructions into Policy-as-Code

arXiv · Research · Trust

Models

When the Aggregator Cheats: Data-Free Backdoors in Federated LLM-based QA Systems

arXiv · Research · Models

Jailbreaking for the Average Jane: Choosing Optimal Jailbreaks via Bandit Algorithms for Automatically Enhanced Queries

arXiv · Research · Models

Reward hacking is swamping model intelligence gains

Cursor · Analysis · Models

AI Security

On the Inseparability of Instructions and Data in Shared-Embedding Sequence Models

arXiv · Research · AI Security

Prompt Injection in Automated Résumé Screening with Large Language Models: Single and Multi-Injection Settings

arXiv · Research · AI Security

ShareLock: A Stealthy Multi-Tool Threshold Poisoning Attack Against MCP

arXiv · Research · AI Security

Advisories in this window

  • 44 npm advisories · 42 critical · 1 moderate · 1 low
  • 19 NuGet advisories · 5 high · 14 moderate
  • 10 PyPI advisories · 1 critical · 2 high · 7 moderate
  • 6 Maven advisories · 4 high · 1 moderate · 1 low
  • 6 Go advisories · 3 high · 3 moderate
  • 2 Packagist advisories · 2 high
  • 2 crates.io advisories · 1 moderate · 1 low

All advisories

Titles, links and short excerpts only; every item links to its source. Selection is automatic; the editor's note is AI-drafted and checked as described. Preview: the unsubscribe link and Keryn's postal address are added to each email when it is sent.