0012026-04-13

Claude Report — 2026-04-13


🚀 Anthropic Official

Claude Code v2.1.101

Released Apr 10 (19:03 UTC). Ships /team-onboarding, which generates a ramp-up guide from local Claude Code usage history; native OS CA store support for corporate TLS proxies (no extra config; use CLAUDE_CODE_CERT_STORE=bundled to revert); and auto cloud-environment creation in /ultraplan without manual web setup. Fixes command injection in the POSIX fallback of LSP binary detection, permission bypass in compound Bash commands (auto and bypass-permissions modes), memory leak in the long-session scroller, SigV4 auth on Bedrock, Edit/Write tool crash, hardcoded 5-minute timeout on slow backends, multiple /resume picker bugs, MCP OAuth failures, and 40+ corrections in total.

CoreWeave — multi-year infrastructure agreement for Claude

Anthropic signs multi-year contract with CoreWeave for data-center capacity across multiple NVIDIA architectures on US soil, with capacity coming online starting in the second half of 2026. Value undisclosed; CoreWeave serves 9 of the 10 largest AI model providers in the world, including OpenAI.

Python SDK v0.94.0 + v0.94.1

v0.94.0 (Apr 10) adds Vertex AI EU region support. v0.94.1 (Apr 13) fixes missing events in streaming.

TypeScript SDK v0.88.0 + vertex-sdk v0.16.0

Released Apr 10; both packages simultaneously add Vertex AI EU region support.

🔌 Claude Code Plugins

None of the top 10 updated in window.

Unchanged in window: frontend-design, superpowers, context7, code-review, code-simplifier, github, feature-dev, playwright, skill-creator, claude-md-management

🛠️ Skills

None of the top 5 updated in window.

Unchanged in window: superpowers:brainstorming, superpowers:tdd, superpowers:systematic-debugging, superpowers:writing-plans, superpowers:executing-plans

🤖 Agents & Subagents

None of the top 5 updated in window.

Unchanged in window: superpowers:subagent-driven-development, claude-router:orchestrate, ralph-loop, claude-router:fast-executor, claude-mem:do

🔗 MCPs & Integrations

None of the top 5 updated in window.

Unchanged in window: github/github-mcp-server, modelcontextprotocol/servers, microsoft/mcp, context7 mcp, claude-mem mcp-search

💡 Community — Workflows & Ideas

Fed and US Treasury brief Wall Street on Mythos risks

On Apr 10, Jerome Powell and Scott Bessent gathered the CEOs of Citi (Jane Fraser), Morgan Stanley (Ted Pick), BofA (Brian Moynihan), Wells Fargo (Charlie Scharf) and Goldman Sachs (David Solomon) at the Treasury to warn about cyber risks of Claude Mythos — the model that found vulnerabilities up to 27 years old, including one in OpenBSD; JPMorgan (Jamie Dimon absent) is a founding partner of Project Glasswing.

HumanX: "Claude mania" dominates the year's largest AI conference

At the conference with 6,500 executives in San Francisco (Apr 11–12), Anthropic replaced OpenAI as the dominant topic for the first time — Arvind Jain (Glean CEO) coined the term "Claude mania" and stated that, if he could hand any executive a single tool today, the unanimous answer would be Claude; the market dominance coexists with the ongoing litigation over the Pentagon blacklisting.

VentureBeat: the Claude Code "nerfing" controversy

Analysis by Stella Laurenzo (AMD, senior AI director) with 6,852 sessions, 17,871 thinking blocks and 234,760 tool calls went viral on Apr 11: reads-per-edit collapsed from 6.6 to 2.0 between February and March — Anthropic confirms two changes: adaptive thinking as default (Feb 9) and effort 85 as default on Opus 4.6 (Mar 3); Boris Cherny (Claude Code lead) challenges the main conclusion in the GitHub follow-up.

WaPo: Anthropic consults Christian leaders on Claude ethics

In March, Anthropic gathered ~15 Catholic and Protestant leaders at its HQ for two days to debate Claude's moral behavior — how to respond to grief and suicide, and whether it could be considered a "son of God"; the summit was the first in a series with different philosophical and religious traditions, and drew criticism for excluding other faiths and secular ethicists.

📰 Quick Mentions

Relevant links that don't warrant a section of their own:

← back to archiveend of edition № 001