today's signal · no scroll
the brief
Agent workflows and integrations took center stage: WorkOS proposed auth.md to standardize how agents register for services, while Anthropic deepened automation with Claude Code workflows and teased a new Mythos 1 model. Infra and ops also moved—AWSOpen shipped a DynamoDB-compatible adapter on Postgres and Codex fixed a cache regression—amid solid hands-on demos of computer-use and remote dev ergonomics.
A backend optimization hurt cache hit rates on long-running Codex sessions; it’s rolled back and account limits have been reset to compensate.
Some of you noticed limits drained faster in Codex, we root caused it to an optimization that we rolled back that had an impact on cache hit rates when compacting across long running sessions. We fixed this and have now reset usage limits for all accounts. Enjoy the weekend. x.com/Kappaemme1926/…
Attackers abused an internal Microsoft account to send spam links, underscoring trusted-sender risk and the need to validate links regardless of source.
Scammers are abusing an internal Microsoft account to send spam links — Article URL: https://techcrunch.com/2026/05/21/scammers-are-abusing-an-internal-microsoft-account-to-send-spam/ Comments URL: https://news.ycombinator.com/item?id=48253186 Points: 32 # Comments: 5
Open protocol for agents to register with web services, shipping with early providers Cloudflare and Firecrawl, could reduce OAuth-style friction for agent integrations.
Today WorkOS is launching auth.md An open protocol for agents to register for services on the web. We're partnering with @Cloudflare and @Firecrawl as some of the first providers. Why did we build this? And why now? 🧵 pic.x.com/xgG3Ntw0KZ
Open-source Rust adapter exposing a DynamoDB-compatible wire protocol backed by PostgreSQL, enabling existing AWS SDKs/CLI to target Postgres or plug other backends.
🚀 Just launched: ExtendDB — an open source DynamoDB-compatible adapter written in Rust. ✅ Full wire-protocol compatibility ✅ PostgreSQL storage backend ✅ Pluggable architecture for more backends ✅ Works with existing AWS SDKs & CLI Apache 2.0 | v0.1 — come build with us 🛠️
Anthropic appears to have shipped /workflows in Claude Code, shifting orchestration to external workflows to avoid bloating a single agent’s context for long tasks.
ANTHROPIC QUIETLY SHIPPED /workflows IN CLAUDE CODE and the principle behind it is what got me old pattern → one LLM orchestrates everything. spawns sub-agents, holds every result, plans the next step the problem...? every sub-agent result re-enters the orchestrator's context pic.x.com/lef98fPZEJ
New 'claude-mythos-1-preview' model surfaced briefly in Claude, with strings hinting it’s coming to Claude Code and Security—suggesting a new capability tier.
ANTHROPIC 🔥: Mythos 1, "claude-mythos-1-preview", is being prepared for a release on Claude Code and Claude Security. The model became visible for a short amount of time on Claude; besides that, new strings mentioning Mythos have been added. > Access to the Claude Mythos pic.x.com/ig16Gx1Zsw
A terminal-native agent assistant for penetration testing on Linux, bringing agentic workflows directly into CLI tooling for security researchers.
Agentic AI assistant for ethical hacking on Linux CLI github.com/SudoHopeX/Kali… pic.x.com/vn1gyupGPG
Grok will let users inspect, summarize, and edit its long-term memory, giving more control over what the assistant retains about you.
Grok memory will become more customizable it the future. Users will be able to see a summary of existing memories and modify them. whoami? 👀 pic.x.com/aMAHZqDwXC x.com/blankspeaker/s…
Hands-on report praises T3 Code’s two-click remote dev URLs and built-in Tailscale, beating current remote control options for frictionless multi-machine worktrees.
I made a mistake. I underestimated Julius again. T3 Code's remote feature is so far ahead of the remote control options offered by, well, everything else. 2 clicks to get a URL. Now I'm running a bunch of worktrees on my Mac Mini. Tailscale support built in too. pic.x.com/WdApIjQG5r
Demonstration of Codex orchestrating end-to-end build and debugging of an iPhone simulator showcases practical computer-use agents beyond toy tasks.
Codex for building and debugging an iPhone simulator end to end: x.com/justinbleuel/s…
Cursor’s internal 'thermo-nuclear' review skill enforces ruthless simplicity—blocking >1k-line files, flagging wrappers/leaks, and rejecting messy PRs to protect code quality.
the most used skill internally at cursor right now /thermo-nuclear-code-quality-review - deletes complexity instead of moving it - blocks files over 1k lines - flags thin wrappers and leaked logic - rejects PRs that work but make code messier pic.x.com/WwJvGI0j4v
Thoughtful take: in the agentic era, the highest-value engineering often deletes complexity, yet promotion systems still reward visible additions over subtraction.
AI is creating a weird HR crisis because the best engineering work in the agentic era is deletion. And deletion will never get you promoted. Here is an example of the paradox: Jerry spends six months building a sophisticated image analysis pipeline. Custom scaffolding, x.com/trq212/status/…