Agent traffic is now heavy enough that the plumbing is being rebuilt around it, as OpenAI swapped its git server, Anthropic added a beta that stops agents resending whole conversations, and Revolut moved coding inference onto its own GPUs.

Agent traffic is now heavy enough that the plumbing is being rebuilt around it, as OpenAI swapped its git server, Anthropic added a beta that stops agents resending whole conversations, and Revolut moved coding inference onto its own GPUs.

Agents are getting their hands on things that cannot be undone, a network's weights, by one founder's account a living animal, a laptop shell reached from a phone, and the part worth studying is the gate around each irreversible step.

The builder news going into Monday is rulebooks written for machines to read, a shared robots.txt training token Google has committed to and a DeepSeek agent that acts only on allowlists and written rules, while Microsoft's threat report shows attackers running agent frameworks with no rulebook at all.

No new models shipped this weekend, so the news came from the meter and the rulebook around agents, with OpenAI pooling ChatGPT and Codex usage, a user reporting Anthropic cut off a training workflow, and Claude Code adding budget and write limits.

With no major launches over the weekend, the signal came from builders publishing what agents cost and where they stall, while the vendor side moved through unannounced refusals and unshipped kernel fixes.

In three unrelated stories this week the outcome was set by code around the model rather than the model, a stealer that switches providers when one refuses, a ChatGPT takeover built from a bundled math library, and a new coding harness whose author says its permission prompts are not security.

The layer underneath the model is where this week's behavior and risk got decided, from a kernel bug an AI model weaponized through a Go default to a study showing the same tools behave differently as MCP or CLI.

The coding-agent failures disclosed on October 30 sat in the plumbing around agents, an editor's extension marketplace and a deployment platform's secret store, while a new benchmark shows the agents themselves are weakest at knowing when to stop and ask.

NVIDIA's auto-ontology shows the safer way to let agents answer data questions, through a read-only business glossary that checks the…

Claude's message threads cut the wait on long agent turns by letting the API hold the conversation for five minutes, but they change…

T3 Code makes one phone app the remote for every coding agent on your machine, and because new threads start in Full access and an unscoped…

The shared usageai-training token finally separates training from crawling in robots.txt, but vendor-specific lines still win, so the…

Dsh-superagent's useful idea is treating a team's attention as the agent's budget, enforced by an allowlist, written rules and an…

The mouse-surgery thread's lesson for builders is its approval plumbing, a human sign-off or a check the agent cannot override on every…

Unslop treats AI-looking UI as a predictable set of failures, so it audits first, stops for a one-screen direction, then fixes nine passes…

The numbat leak shows a frontier label can collect thousands of downloads with no provenance, and the real danger in trying such a model is…