Click any tag below to further narrow down your results
+ open source
(3)
+ anthropic
(2)
+ gemini 4 argon
(2)
+ code generation
(2)
+ vulnerability scanning
(1)
+ system prompts
(1)
+ evaluation
(1)
+ bug bounty programs
(1)
+ ai submissions
(1)
+ ai agent security
(1)
+ splunk agentic soc
(1)
+ widefield security
(1)
+ prompt injection
(1)
+ claude bootstrap
(1)
+ agent security
(1)
Links
OpenAI is pitching agents that operate EHRs and payer portals by clicking and typing, so they can handle tasks like MRI authorization that APIs don't reach. The post shares evaluation results on helpfulness, safety, and cost, plus early deployments with Cigna, CVS Health, Abridge, and Ambience.
- A demo agent takes a patient's message and visit notes, pulls prior imaging and PT records from the EHR, and drafts an MRI order, encounter note, and patient reply for clinician review.
- CVS Health reports an agent cleared a months-long document-sending backlog over a weekend and handles system steps for tens of thousands of consults a day, with validation checks and human oversight.
- Cigna case managers previously had to navigate more than 10 separate systems to gather member context; computer use now pulls that together in seconds.
Google is rolling out Gemini 4 Carbon, its most capable model, aimed at hard tasks like multi-step coding, scientific research, and cybersecurity work. It costs more to run than the lighter Gemini 4 Argon, but Google says it's cheaper per task than competing top models on several benchmarks.
- Priced at $2 per million input tokens and $10 per million output tokens, the same introductory rate as Argon, and Google claims more than 2x lower cost per task than the next best external model on CursorBench 4.0, WANDR, and AutomationBench.
- Google previews /swarm in Google Antigravity, which coordinates multiple agents with Carbon handling architecture and auditing while lighter models do parallel work.
- In cybersecurity, Google reports Carbon found 0-day vulnerabilities in two major SaaS providers in under two hours, though these claims come from Google and its partners rather than independent testing.
Anthropic is offering free, fully model-generated security scans to eligible open source projects through a new program called OSS Scanner, with reports delivered without human triage. Early testers say the findings were mostly valid, though the company admits some reports will be wrong.
- Anthropic's models found over 29,000 candidate vulnerabilities in six months, but only about 6,000 have been manually reviewed, so human review is the bottleneck.
- OSS Scanner reports go out unverified, but in a pilot, 85 of 97 critical and high-severity findings (88%) met the bar for Anthropic's disclosure process, and only one was a false positive.
- Eligible projects get reports with a reproducer, an explanation, and often a candidate patch; several maintainers said the reports were high quality, and some chained bugs into unauthenticated remote code execution.
Google released Gemini 4 Argon, a new frontier model designed to handle complex multi-step business tasks across software development, finance, legal work, and security. The model can maintain context over long documents, deliver precise results with verifiable citations, and execute autonomous work from a single prompt.
- Argon handles multi-step tasks that previously took hours or days, processing long documents with mixed text, tables, charts, and images into structured outputs with honest reporting of limitations.
- In regulated domains, Argon leads on benchmarks for precision: it tops the Vals Index for financial and legal work, scores 68% on CWE-Bench for security vulnerability remediation, and achieved 20% better task success than earlier models in ServiceNow's testing.
- Google's internal teams used Argon to cut code session prep time by 40%, reduce rollbacks by nearly a third, and compress months of architectural refactoring work into weeks on Chrome's 40-million-line codebase.
- Argon resists indirect prompt injection attacks at 0.7% success rate across 15 adaptive attempts and includes defensive security capabilities while blocking offensive tasks like pentesting.
Coinbase built an LLM-powered fraud detection agent for its Onramp crypto payment service that sits downstream of existing ML models and rules, reviewing transaction sequences for patterns that emerge across multiple purchases. In live testing, the agent reduced fraudulent transactions by 30% and fraud value by 22%, with projected fraud prevention savings roughly three times the inference cost.
- The agent catches fraud patterns invisible in single transactions by reviewing structured context, historical summaries, and transaction sequences together—something the existing ML model and rules engine miss.
- Selective routing of transactions to the LLM balances fraud detection against latency and cost, with the agent only overriding decisions to block (never to allow) what the upstream checks flagged.
- Newer model versions don't automatically improve domain-specific performance: Part 2 found that newer Opus, Sonnet, and GPT releases all scored lower on recall and F1 for fraud detection, motivating Coinbase to post-train an open-weight model instead.
This is a news digest from The Register covering major developments across AI, infrastructure, security, and open source — including TSMC's US expansion, Microsoft's AI agent plans, Russian phishing campaigns, and various software releases.
- TSMC is partnering with GlobalFoundries on a $2B deal to boost US production of silicon interposers needed for AI datacenter chips
- Microsoft is integrating Copilot agents into Windows file systems and replacing Intel processors with Nvidia SoCs in Surface Laptop Ultra, starting at $2,599
- Russian attackers are impersonating Signal support to run phishing campaigns, and Microsoft's SharePoint on-premises servers are under active zero-day attack with unpatched vulnerabilities
- Ubuntu's new beta adds Rust components and LLM-based speech-to-text; Firefox 157 gets a significant visual redesign; KDE turns 30 with an AI-native desktop proposal
Google paused its open-source bug bounty after AI flooded it with fake vulnerability reports, forcing security teams to waste time triaging hallucinated exploits. Nulsec proposes an automated sandbox that pre-screens submissions to filter out AI slop before human engineers see them.
- Google froze its open-source bug bounty program because AI-generated reports created an unsustainable triage load—most submissions describe non-existent vulnerabilities or non-reproducible exploits, wasting hundreds of engineering hours.
- Nulsec's core mechanism: automatically execute reproduction tests in isolated sandboxes for each incoming report, blocking hallucinated claims before they reach human security engineers.
- The business model stacks three revenue streams (tiered SaaS, usage-based compute, enterprise integrations) and uses free tier access to top GitHub repos as a go-to-market wedge.
Google froze its open source bug bounty program starting October 1 because AI-generated submissions overwhelmed the system with invalid reports and hallucinations. The company won't reopen it until Q1 2027.
- Google engineers and open source maintainers were flooded with automated submissions that were mostly invalid or contained AI hallucinations, making the program unmanageable.
- The pause confirms cybersecurity experts' earlier warnings that AI-generated content poses a real threat to bug bounty programs by degrading their signal-to-noise ratio.
- Google is redirecting participants to its other bug bounty programs while it figures out how to handle the AI submission problem.
Cisco will buy WideField Security and fold its identity and session telemetry tech into Splunk’s agentic AI SOC. The deal aims to track human and AI-agent actions in real time, tightening security around automated workflows. It follows Cisco’s recent picks of Astrix Security and Galileo Technologies.
- Cisco is acquiring WideField Security (fresh off an $11.3M Series A) to add real-time identity/session tracking for humans and AI agents into Splunk's agentic SOC.
- This is Cisco's third AI-security acquisition in recent months, following Astrix Security (agent protection) and Galileo Technologies (Splunk Observability).
- The tech aims to catch threats invisible to legacy tools, like authenticated AI processes going rogue, by feeding session-level signals into Cisco's Data Fabric.
- Customer demand for this isn't strong yet, but partners see deep identity tracking becoming essential as AI agents proliferate.
SkillSpector scans AI agent skills for vulnerabilities, malicious patterns, and CVEs before installation. It combines fast static checks with optional LLM analysis, covers 64 patterns across 16 categories, and outputs risk scores and detailed reports.
- SkillSpector scans AI agent skills before installation, checking 64 vulnerability patterns across 16 categories (prompt injection, data exfiltration, privilege escalation, etc.) plus real-time CVE lookups via OSV.dev.
- It runs a fast static scan by default, with an optional deeper LLM-based semantic check using providers like OpenAI, Anthropic, or NVIDIA's build.nvidia.com.
- Scans output a 0-100 risk score with severity labels and suggested fixes, in formats including JSON, Markdown, or SARIF for CI/CD pipelines.
- Even without LLM analysis (--no-llm flag), it still catches known code patterns and CVEs.
Anthropic is holding back its new AI model, Claude Mythos Preview, and teaming up with over 40 tech firms to hunt and patch security flaws in critical software. The company says the model can autonomously find zero-day vulnerabilities that have eluded researchers for decades, raising fresh concerns about AI-driven cyberattacks.
- Anthropic is withholding public release of Claude Mythos Preview and instead giving early access to ~40 companies (Apple, Amazon, Microsoft, Google, Cisco, Broadcom, Linux Foundation) under "Project Glasswing," backed by up to $100 million in usage credits, to find and patch critical software vulnerabilities first.
- The model has reportedly found a 27-year-old vulnerability in OpenBSD and a flaw in video software that survived five million automated scans, using simple prompts to autonomously hunt zero-days.
- Anthropic frames this as a security "reckoning" while simultaneously racing toward projected revenue of $30 billion this year, mirroring the tension of building powerful AI it also warns could enable dangerous cyberattacks.
Claude Bootstrap is an opinionated system designed for initializing projects with a focus on test-driven development (TDD), security, and simplicity. It automates iterative coding loops, ensures mandatory code reviews, and helps maintain clarity and security in AI-generated code. The framework encapsulates best practices learned from numerous AI-assisted projects across various programming environments.
- Bakes TDD, security checks, and mandatory code review directly into Claude Code's project scaffolding rather than leaving them to developer discipline
- Automates iterative coding loops so AI-generated code gets continuously tested and refined instead of accepted on first pass
- Distills lessons from many prior AI-assisted projects into one opinionated, reusable initialization framework across languages/environments
User feedback is highly valued and taken seriously by the team. For a comprehensive list of available qualifiers, users are directed to the documentation. An error occurred while loading the page, prompting a reload.
- The page failed to load properly, so the actual leaked Anthropic system prompt content isn't accessible in this summary
- The repository (x1xhlol/system-prompts-and-models-of-ai-tools) compiles system prompts from AI tools, positioning it as a reference resource rather than original commentary