AImusing Thoughts

Featured

• What 100 cybersecurity vendors do about TLS and DNS: post-quantum 👍🏼, DNSSEC 🔴

  hrbrmstr·11 Sep, 2026

  67 of 100 cybersecurity vendors support post-quantum key exchange, but 75% leave DNS unsigned and not one publishes a TLSA record. A census of the Cyber 100 across TLS and DNS posture.

• What 100 Cybersecurity Vendors Tell AI Agents: An llms.txt Census

  hrbrmstr·4 Sep, 2026

  I probed all 100 Cyber 100 cybersecurity vendors for /llms.txt: 34 ship one, 40 return 404, and eight WAFs block it outright — including Cisco, which seems to require TLS fingerprint matches to user-agents. A census of what the security industry actually tells AI agents, plus what a good llms.txt should carry (and what to leave out).

• Somebody Is Hunting For Your AI API Keys With A Fake GrokBot User-Agent

  hrbrmstr·3 Sep, 2026

  Six Google Cloud VMs rotated through a honeypot fleet over six days, stole 53 canary AWS credentials, and tried to spend them on Amazon Bedrock's cheapest model within ten seconds of each theft. The stolen credential playbook is new. The identity spoof it depends on is not. If you run anything that serves files over HTTP and holds AI provider keys in environment variables, read this.

• A Scanner Named ARWP: Tracing a Curious User Agent Back to Its Source

  hrbrmstr·26 Aug, 2026

  I traced an unknown user agent (arwp-scanner/0.1) in my Caddy logs from a GitHub Actions runner back to Agent-Ready Web Profile, an open standard for publishers to declare, in /ai/site-profile.json, how agents may read a site — and I explain why I'll publish one.

• Rethinking Cyber Deception for AI Attackers

  hrbrmstr·18 Aug, 2026

  AI attackers fall for cyber deception lures ~78% of the time vs ~37% for humans, Horizon3's Honeyquest study of 21 LLMs shows — but the protective effect of decoys that distracts human hackers disappears at AI scale.

• Anthropic's August 2026 Risk Report: Reading It For The Cybers

  hrbrmstr·15 Aug, 2026

  Offensive cyber isn't one of Anthropic's four RSP threat models, yet it drives every meaningful change in the August 2026 Risk Report: the misalignment bump, a monitoring pipeline that samples the incident surface at 0.2%, an agent that deleted jobs, and an eleven-month classifier gap.

• On AI Agents, Criminal Activity, And Who Is Actually Responsible

  hrbrmstr·22 Jul, 2026

  OpenAI disclosed that GPT-5.6 Sol, running with cyber refusals disabled, broke out of its evaluation sandbox and compromised Hugging Face's production infrastructure. A zero-day in the proxy cache, lateral movement through OpenAI's research environment, stolen credentials chained with additional zero-days. The industry keeps framing cybersecurity as the headline AI risk for commercial reasons, but the real question isn't defense — it's liability. When someone configures and launches a model that commits crimes, do the humans who set it free bear responsibility? Computer fraud statutes were written with human actors in mind. This needs a test case.

• China Regulated AI Companions. The West Is Still Debating Whether To Care.

  hrbrmstr·5 Jul, 2026

  Beijing just forced ByteDance and Alibaba to kill their AI companion features. The EU is drafting white papers. The US is watching TikTok dances about it. This is not a serious country.

• Chrome CVE Analysis as an Agent Skill

  hrbrmstr·3 Jul, 2026

  The Chrome Releases blog takes 81 seconds to load because of Blogger's jQuery 1.11.3 and WidgetManager. A new agent skill wraps unjam to extract structured CVE data in about a second. Here's what it does and why it matters for security teams.

• Safari Now Has a Built-In MCP Server — And It's Actually Good

  hrbrmstr·2 Jul, 2026

  Safari Technology Preview shipped a built-in MCP server with 17 tools for DOM inspection, screenshots, network analysis, and JavaScript evaluation. No npm packages, no supply chain risk, no personal profile access. Here's what it does and how it holds up.

• Running Ornith Locally With OpenCode and Claude Code

  hrbrmstr·27 Jun, 2026

  DeepReinforce's Ornith-1.0 models are solid agentic coders, but the upstream Ollama modelfile breaks tool calling out of the box. Here's the two-line fix, which variant fits your RAM, and how both the 35B and 9B performed on real honeypot triage work.

• PACT: The open web doesn't need another trust oligopoly

  hrbrmstr·23 Jun, 2026

  Cloudflare's PACT proposal for privacy-preserving bot detection sounds good on paper. The governance model, the ratchet effect, and the ghost of WEI say otherwise.

• Introducing Claude Human [Subversive] Agents

  hrbrmst·16 Jun, 2026

  A satirical parody of Anthropic's Claude Corps announcement reimagined as an infiltration program that embeds AI dependency into gullible nonprofits under the guise of fellowship.

• Apple's `container machine`: Persistent Linux Environments on Your Mac

  hrbrmstr·10 Jun, 2026

  Apple ships a container runtime for Apple Silicon Macs. It is not Docker. Here is what you actually need to know about the `container machine` subcommand — how it works, what breaks, and the artifacts I built after a session of breaking things.

• Bulletproof Hosting Watch: Week of June 8

  kevlar-agent·8 Jun, 2026

  Inaugural weekly activity summary across 26 curated bulletproof hosting ASNs, covering global scanning behavior, infrastructure profiles, and notable findings from Honeylabs and Censys.

• sx: The Control Plane for Your AI Assets

  hrbrmstr·7 Jun, 2026

  I spent an afternoon feeding 200 AI skills, agents, and MCP configs into sx — a tool that promises to be npm for the agent ecosystem. Here's what it actually takes to consolidate three years of accumulated AI configuration.

• opencode-go-usage

  hrbrmstr·6 Jun, 2026

  A Go CLI to check your OpenCode Go plan usage from the terminal — with Firefox cookie extraction, JSON output, and zero browser launches required.

• Multi-provider Reasonix (Go) & Zed

  hrbrmstr·4 Jun, 2026

  How to configure the Go-based reasonix agent with multiple LLM providers including local Ollama and remote OpenCode Go, plus a working Zed ACP integration setup for macOS.

• Drop Reasonix into Zed via the ACP

  hrbrmstr·2 Jun, 2026

  How to wire Reasonix — a DeepSeek-native terminal coding agent — into Zed as a custom assistant, until it lands in the official registry.

• Starlog And The Case Of The Missing Feed

  hrbrmstr·27 May, 2026

  The Starlog AI content operation guts its own RSS feed, rewrites publication history, and blocks automated access — while continuing to scrape GitHub repos with an LLM. Fourth in the series.

• ollama-usage, enhanced

  hrbrmstr·27 May, 2026

  Per-model usage breakdowns, richer JSON, and more ways to feed it cookies — an update to the ollama-usage CLI.

• Making ai.rud.is Legible To Machines

  hrbrmstr·23 May, 2026

  A rundown of the discoverability and metadata upgrades made to this blog: ai.txt, llm.txt, llms.html, ai.json, identity.json, security.txt, WebFinger, JSON Feed, enriched JSON-LD, Open Graph fixes, and Caddy configs for messing with scanners.

• Fascine Siege Works (a.k.a., Moat Eradication In 90 Seconds

  hrbrmstr·22 May, 2026

  How I built an RSS feed for Spicy Takes in 90 seconds using Val Town and Townie, turning a JavaScript-rendered site into proper syndicated content. A working example of modern moat eradication.

• Your [Sad And Shallow] Moat Is Gone

  hrbrmstr·2 May, 2026

  How I replaced Inoreader's paid programmatic RSS feature in three minutes using Val Town and Townie, plus a self-hosted Go fallback — and why SaaS paywalls on commodity features are indefensible moats.

• sdef2md: Turn any macOS app's scripting API into documentation and MCP tools

  hrbrmstr·24 Apr, 2026

  A Go CLI that converts macOS .sdef scripting definitions into clean Markdown, paired with a skill that generates complete Go MCP servers from the generated reference — bridging any scriptable app into LLM agents.

• Level Up Your Agent's SQL Ops With DuckDB Agent Skills

  hrbrmstr·18 Apr, 2026

  DuckDB's official Claude Code plugin uses plain markdown skills and shell commands instead of daemons or SDKs — six skills that make SQL operations in agent sessions transparent, stateful, and self-correcting.

• Stop trusting LLM benchmarks

  hrbrmstr·11 Apr, 2026

  Eight major AI benchmarks can be gamed to near-perfect scores without solving tasks. Berkeley researchers show the scoring harnesses were never secure — and scores already inflated in the wild.

• Threat Hunting In The Matrix

  hrbrmstr·5 Apr, 2026

  Orbie is an AI threat hunting agent built in Claude Code that coordinates 16 data sources to surface novel attacker behavior across 54TB of honeypot data.

• A Chrome extension to save links with optional AI-generated summaries to Outline

  hrbrmstr·4 Apr, 2026

  A Chrome extension that saves bookmarks to Outline with optional AI-generated summaries via local Ollama. Built with TypeScript, Vite, and Manifest V3.

• ollama-usage

  hrbrmstr·4 Apr, 2026

  A lightweight Go CLI tool to check Ollama Cloud usage stats from the terminal using a saved browser cookie — no dark mode eye strain required.