vibehacker
News
Arcjet Blog ·

Arcjet ships runtime security for coding agents across Claude Code, Codex, Cursor, and Copilot

Announced Oct 8, Arcjet Runtime Security for Coding Agents discovers agents on developer endpoints and applies one Rego policy set across Claude Code, Codex, Cursor, Muse Code, and Copilot through each agent's hooks, with detectors for prompt injection, PII, and high-risk MCP or web destinations. It's available now; Arcjet recommends publishing the starter policies in dry-run for a week before enforcing.

Read at arcjet.com

More news

View all

Anthropic launches OSS Scanner, free unreviewed Claude security scans for critical open-source projects

Launched Oct 8 under the new Anthropic Cyber Mission, OSS Scanner gives enrolled projects periodic scans from Anthropic's strongest models, with a proof of concept and suggested fix per bug, sent without human review (Anthropic expects a true positive rate above 90%). It's opt in for established projects with critical infrastructure or user security impact, and maintainers enroll by opening a PR to the anthropics/oss scanner repo…

Anthropic

Anthropic launches Claude Dashboards and Claude Motion in beta; Docs, Slides, and Design go GA

Announced Oct 8, Claude Dashboards (paid plans) connects to Databricks, Snowflake, Redshift, ClickHouse, or Salesforce and builds auto refreshing dashboards from plain language questions with a clickable query behind every number; Claude Motion (Team and Enterprise) writes code based editable animations you can export as MP4. Claude Docs, Slides, and Design left beta the same day and are now on every plan, including Free…

Digital Trends

Claude Code 2.1.295 adds onFailure:block for hooks and Program Status Protocol for terminals

Released Oct 8, command and HTTP hooks can set onFailure: "block" so a hook that can't start, times out, or exits unexpectedly blocks the action instead of letting it through, and terminals that speak OSC 7501 can show whether Claude Code is working, waiting on you, or done. It also raises the MCP tool description cap through tool search from 2,048 to 16,384 characters, backs off remote MCP reconnects after outages, and folds in 2.1.294's fix for prompt/agent hooks that were allowing what they were written to block…

GitHub / anthropics/claude-code releases

Functionize ships an MCP server so Claude Code, Cursor, and Copilot can build and run app tests from chat

Released Oct 8, the Functionize MCP server lets Claude Code, Cursor, GitHub Copilot, and other MCP clients ask Functionize Studio in plain language to build and run a smoke or regression suite against the real app, then get the failing step back in the same chat. It's available now to Studio customers and trial users, with a Claude Code plugin that adds skills for Studio…

Functionize Blog

Codex Day 4: instant steering ships alongside GPT-6.1 Sol Ultrafast, up to 8x faster than Standard

Codex lead Tibo's Day 4 post (Oct 8) says mid task steering now takes effect instantly, so the model reacts to your corrections right away instead of wasting effort on the old direction, while OpenAI Developers says Ultrafast for GPT 6.1 Sol is rolling out today in the API, Codex, and ChatGPT Work at up to 8x Sol Standard's speed. Astra's Ultrafast has been limited to Pro $500 and eligible Enterprise and Edu plans and bills above Standard, so check your plan and the API pricing page before switching it on…

X / @thsottiaux (Codex lead)

Spotted something we missed? Start a thread.